Authors: Junggu Choi, Tak Hur, Daniel K. Park, Na-Young Shin, Seung-Koo Lee, Hakbae Lee, Sanghoon Han
Abstract: Following the recent development of quantum machine learning techniques, the literature has reported several quantum machine learning algorithms for disease detection. This study explores the application of a hybrid quantum-classical algorithm for classifying region-of-interest time-series data obtained from resting-state functional magnetic resonance imaging in patients with early-stage cognitive impairment based on the importance of cognitive decline for dementia or aging. Classical one-dimensional convolutional layers are used together with quantum convolutional neural networks in our hybrid algorithm. In the classical simulation, the proposed hybrid algorithms showed higher balanced accuracies than classical convolutional neural networks under the similar training conditions. Moreover, a total of nine brain regions (left precentral gyrus, right superior temporal gyrus, left rolandic operculum, right rolandic operculum, left parahippocampus, right hippocampus, left medial frontal gyrus, right cerebellum crus, and cerebellar vermis) among 116 brain regions were found to be relatively effective brain regions for the classification based on the model performances. The associations of the selected nine regions with cognitive decline, as found in previous studies, were additionally validated through seed-based functional connectivity analysis. We confirmed both the improvement of model performance with the quantum convolutional neural network and neuroscientific validities of brain regions from our hybrid quantum-classical model.
Authors: Mustafa Cavus, Przemys{\l}aw Biecek
Abstract: Predictive models may generate biased predictions when classifying imbalanced datasets. This happens when the model favors the majority class, leading to low performance in accurately predicting the minority class. To address this issue, balancing or resampling methods are critical pre-processing steps in the modeling process. However, there have been debates and questioning of the functionality of these methods in recent years. In particular, many candidate models may exhibit very similar predictive performance, which is called the Rashomon effect, in model selection. Selecting one of them without considering predictive multiplicity which is the case of yielding conflicting models' predictions for any sample may lead to a loss of using another model. In this study, in addition to the existing debates, the impact of balancing methods on predictive multiplicity is examined through the Rashomon effect. It is important because the blind model selection is risky from a set of approximately equally accurate models. This may lead to serious problems in model selection, validation, and explanation. To tackle this matter, we conducted real dataset experiments to observe the impact of balancing methods on predictive multiplicity through the Rashomon effect. Our findings showed that balancing methods inflate the predictive multiplicity, and they yield varying results. To monitor the trade-off between performance and predictive multiplicity for conducting the modeling process responsibly, we proposed using the extended performance-gain plot for the Rashomon effect.
Authors: Yasin Abbasi Yadkori, Ilja Kuzborskij, David Stutz, Andr\'as Gy\"orgy, Adam Fisch, Arnaud Doucet, Iuliya Beloshapka, Wei-Hung Weng, Yao-Yuan Yang, Csaba Szepesv\'ari, Ali Taylan Cemgil, Nenad Tomasev
Abstract: We develop a principled procedure for determining when a large language model (LLM) should abstain from responding (e.g., by saying "I don't know") in a general domain, instead of resorting to possibly "hallucinating" a non-sensical or incorrect answer. Building on earlier approaches that use self-consistency as a more reliable measure of model confidence, we propose using the LLM itself to self-evaluate the similarity between each of its sampled responses for a given query. We then further leverage conformal prediction techniques to develop an abstention procedure that benefits from rigorous theoretical guarantees on the hallucination rate (error rate). Experimentally, our resulting conformal abstention method reliably bounds the hallucination rate on various closed-book, open-domain generative question answering datasets, while also maintaining a significantly less conservative abstention rate on a dataset with long responses (Temporal Sequences) compared to baselines using log-probability scores to quantify uncertainty, while achieveing comparable performance on a dataset with short answers (TriviaQA). To evaluate the experiments automatically, one needs to determine if two responses are equivalent given a question. Following standard practice, we use a thresholded similarity function to determine if two responses match, but also provide a method for calibrating the threshold based on conformal prediction, with theoretical guarantees on the accuracy of the match prediction, which might be of independent interest.
Authors: Yunhui Guo
Abstract: Transferability estimation has emerged as an important problem in transfer learning. A transferability estimation method takes as inputs a set of pre-trained models and decides which pre-trained model can deliver the best transfer learning performance. Existing methods tackle this problem by analyzing the output of the pre-trained model or by comparing the pre-trained model with a probe model trained on the target dataset. However, neither is sufficient to provide reliable and efficient transferability estimations. In this paper, we present a novel perspective and introduce Kite, as a Kernel-based Improved Transferability Estimation method. Kite is based on the key observations that the separability of the pre-trained features and the similarity of the pre-trained features to random features are two important factors for estimating transferability. Inspired by kernel methods, Kite adopts centered kernel alignment as an effective way to assess feature separability and feature similarity. Kite is easy to interpret, fast to compute, and robust to the target dataset size. We evaluate the performance of Kite on a recently introduced large-scale model selection benchmark. The benchmark contains 8 source dataset, 6 target datasets and 4 architectures with a total of 32 pre-trained models. Extensive results show that Kite outperforms existing methods by a large margin for transferability estimation.
Authors: Zhengsen Xu, Jonathan Li, Linlin Xu
Abstract: Wildfires have significant impacts on global vegetation, wildlife, and humans. They destroy plant communities and wildlife habitats and contribute to increased emissions of carbon dioxide, nitrogen oxides, methane, and other pollutants. The prediction of wildfires relies on various independent variables combined with regression or machine learning methods. In this technical review, we describe the options for independent variables, data processing techniques, models, independent variables collinearity and importance estimation methods, and model performance evaluation metrics. First, we divide the independent variables into 4 aspects, including climate and meteorology conditions, socio-economical factors, terrain and hydrological features, and wildfire historical records. Second, preprocessing methods are described for different magnitudes, different spatial-temporal resolutions, and different formats of data. Third, the collinearity and importance evaluation methods of independent variables are also considered. Fourth, we discuss the application of statistical models, traditional machine learning models, and deep learning models in wildfire risk prediction. In this subsection, compared with other reviews, this manuscript particularly discusses the evaluation metrics and recent advancements in deep learning methods. Lastly, addressing the limitations of current research, this paper emphasizes the need for more effective deep learning time series forecasting algorithms, the utilization of three-dimensional data including ground and trunk fuel, extraction of more accurate historical fire point data, and improved model evaluation metrics.
Authors: Benjamin Sykes, Loic Simon, Julien Rabin
Abstract: With the recent success of generative models in image and text, the evaluation of generative models has gained a lot of attention. Whereas most generative models are compared in terms of scalar values such as Frechet Inception Distance (FID) or Inception Score (IS), in the last years (Sajjadi et al., 2018) proposed a definition of precision-recall curve to characterize the closeness of two distributions. Since then, various approaches to precision and recall have seen the light (Kynkaanniemi et al., 2019; Naeem et al., 2020; Park & Kim, 2023). They center their attention on the extreme values of precision and recall, but apart from this fact, their ties are elusive. In this paper, we unify most of these approaches under the same umbrella, relying on the work of (Simon et al., 2019). Doing so, we were able not only to recover entire curves, but also to expose the sources of the accounted pitfalls of the concerned metrics. We also provide consistency results that go well beyond the ones presented in the corresponding literature. Last, we study the different behaviors of the curves obtained experimentally.
Authors: Christian Marius Lillelund, Fernando Pannullo, Morten Opprud Jakobsen, Manuel Morante, Christian Fischer Pedersen
Abstract: Predicting the remaining useful life (RUL) of ball bearings plays an important role in predictive maintenance. A common definition of the RUL is the time until a bearing is no longer functional, which we denote as an event, and many data-driven methods have been proposed to predict the RUL. However, few studies have addressed the problem of censored data, where this event of interest is not observed, and simply ignoring these observations can lead to an overestimation of the failure risk. In this paper, we propose a probabilistic estimation of RUL using survival analysis that supports censored data. First, we analyze sensor readings from ball bearings in the frequency domain and annotate when a bearing starts to deteriorate by calculating the Kullback-Leibler (KL) divergence between the probability density function (PDF) of the current process and a reference PDF. Second, we train several survival models on the annotated bearing dataset, capable of predicting the RUL over a finite time horizon using the survival function. This function is guaranteed to be strictly monotonically decreasing and is an intuitive estimation of the remaining lifetime. We demonstrate our approach in the XJTU-SY dataset using cross-validation and find that Random Survival Forests consistently outperforms both non-neural networks and neural networks in terms of the mean absolute error (MAE). Our work encourages the inclusion of censored data in predictive maintenance models and highlights the unique advantages that survival analysis offers when it comes to probabilistic RUL estimation and early fault detection.
Authors: Lena Todnem Bach Christensen, Dikte Straadt, Stratos Vassis, Christian Marius Lillelund, Peter Bangsgaard Stoustrup, Ruben Pauwels, Thomas Klit Pedersen, Christian Fischer Pedersen
Abstract: Juvenile idiopathic arthritis (JIA) is the most common rheumatic disease during childhood and adolescence. The temporomandibular joints (TMJ) are among the most frequently affected joints in patients with JIA, and mandibular growth is especially vulnerable to arthritic changes of the TMJ in children. A clinical examination is the most cost-effective method to diagnose TMJ involvement, but clinicians find it difficult to interpret and inaccurate when used only on clinical examinations. This study implemented an explainable artificial intelligence (AI) model that can help clinicians assess TMJ involvement. The classification model was trained using Random Forest on 6154 clinical examinations of 1035 pediatric patients (67% female, 33% male) and evaluated on its ability to correctly classify TMJ involvement or not on a separate test set. Most notably, the results show that the model can classify patients within two years of their first examination as having TMJ involvement with a precision of 0.86 and a sensitivity of 0.7. The results show promise for an AI model in the assessment of TMJ involvement in children and as a decision support tool.
Authors: Bettina Finzel, Patrick Hilme, Johannes Rabold, Ute Schmid
Abstract: Explanations for Convolutional Neural Networks (CNNs) based on relevance of input pixels might be too unspecific to evaluate which and how input features impact model decisions. Especially in complex real-world domains like biomedicine, the presence of specific concepts (e.g., a certain type of cell) and of relations between concepts (e.g., one cell type is next to another) might be discriminative between classes (e.g., different types of tissue). Pixel relevance is not expressive enough to convey this type of information. In consequence, model evaluation is limited and relevant aspects present in the data and influencing the model decisions might be overlooked. This work presents a novel method to explain and evaluate CNN models, which uses a concept- and relation-based explainer (CoReX). It explains the predictive behavior of a model on a set of images by masking (ir-)relevant concepts from the decision-making process and by constraining relations in a learned interpretable surrogate model. We test our approach with several image data sets and CNN architectures. Results show that CoReX explanations are faithful to the CNN model in terms of predictive outcomes. We further demonstrate that CoReX is a suitable tool for evaluating CNNs supporting identification and re-classification of incorrect or ambiguous classifications.
Authors: Yufei Jin, Xingquan Zhu
Abstract: Oversmoothing is a commonly observed challenge in graph neural network (GNN) learning, where, as layers increase, embedding features learned from GNNs quickly become similar/indistinguishable, making them incapable of differentiating network proximity. A GNN with shallow layer architectures can only learn short-term relation or localized structure information, limiting its power of learning long-term connection, evidenced by their inferior learning performance on heterophilous graphs. Tackling oversmoothing is crucial to harness deep-layer architectures for GNNs. To date, many methods have been proposed to alleviate oversmoothing. The vast difference behind their design principles, combined with graph complications, make it difficult to understand and even compare their difference in tackling the oversmoothing. In this paper, we propose ATNPA, a unified view with five key steps: Augmentation, Transformation, Normalization, Propagation, and Aggregation, to summarize GNN oversmoothing alleviation approaches. We first outline three themes to tackle oversmoothing, and then separate all methods into six categories, followed by detailed reviews of representative methods, including their relation to the ATNPA, and discussion about their niche, strength, and weakness. The review not only draws in-depth understanding of existing methods in the field, but also shows a clear road map for future study.
Authors: Shangding Gu, Bilgehan Sel, Yuhao Ding, Lu Wang, Qingwei Lin, Ming Jin, Alois Knoll
Abstract: Ensuring the safety of Reinforcement Learning (RL) is crucial for its deployment in real-world applications. Nevertheless, managing the trade-off between reward and safety during exploration presents a significant challenge. Improving reward performance through policy adjustments may adversely affect safety performance. In this study, we aim to address this conflicting relation by leveraging the theory of gradient manipulation. Initially, we analyze the conflict between reward and safety gradients. Subsequently, we tackle the balance between reward and safety optimization by proposing a soft switching policy optimization method, for which we provide convergence analysis. Based on our theoretical examination, we provide a safe RL framework to overcome the aforementioned challenge, and we develop a Safety-MuJoCo Benchmark to assess the performance of safe RL algorithms. Finally, we evaluate the effectiveness of our method on the Safety-MuJoCo Benchmark and a popular safe benchmark, Omnisafe. Experimental results demonstrate that our algorithms outperform several state-of-the-art baselines in terms of balancing reward and safety optimization.
Authors: Nima Hosseini Dashtbayaz, Ghazal Farhani, Boyu Wang, Charles X. Ling
Abstract: The residual loss in Physics-Informed Neural Networks (PINNs) alters the simple recursive relation of layers in a feed-forward neural network by applying a differential operator, resulting in a loss landscape that is inherently different from those of common supervised problems. Therefore, relying on the existing theory leads to unjustified design choices and suboptimal performance. In this work, we analyze the residual loss by studying its characteristics at critical points to find the conditions that result in effective training of PINNs. Specifically, we first show that under certain conditions, the residual loss of PINNs can be globally minimized by a wide neural network. Furthermore, our analysis also reveals that an activation function with well-behaved high-order derivatives plays a crucial role in minimizing the residual loss. In particular, to solve a $k$-th order PDE, the $k$-th derivative of the activation function should be bijective. The established theory paves the way for designing and choosing effective activation functions for PINNs and explains why periodic activations have shown promising performance in certain cases. Finally, we verify our findings by conducting a set of experiments on several PDEs. Our code is publicly available at https://github.com/nimahsn/pinns_tf2.
Authors: Darshan Patil, Janarthanan Rajendran, Glen Berseth, Sarath Chandar
Abstract: In the real world, the strong episode resetting mechanisms that are needed to train agents in simulation are unavailable. The \textit{resetting} assumption limits the potential of reinforcement learning in the real world, as providing resets to an agent usually requires the creation of additional handcrafted mechanisms or human interventions. Recent work aims to train agents (\textit{forward}) with learned resets by constructing a second (\textit{backward}) agent that returns the forward agent to the initial state. We find that the termination and timing of the transitions between these two agents are crucial for algorithm success. With this in mind, we create a new algorithm, Reset Free RL with Intelligently Switching Controller (RISC) which intelligently switches between the two agents based on the agent's confidence in achieving its current goal. Our new method achieves state-of-the-art performance on several challenging environments for reset-free RL.
Authors: Simon Vary, Pierre Ablin, Bin Gao, P. -A. Absil
Abstract: Optimization over the set of matrices that satisfy $X^\top B X = I_p$, referred to as the generalized Stiefel manifold, appears in many applications involving sampled covariance matrices such as canonical correlation analysis (CCA), independent component analysis (ICA), and the generalized eigenvalue problem (GEVP). Solving these problems is typically done by iterative methods, such as Riemannian approaches, which require a computationally expensive eigenvalue decomposition involving fully formed $B$. We propose a cheap stochastic iterative method that solves the optimization problem while having access only to a random estimate of the feasible set. Our method does not enforce the constraint in every iteration exactly, but instead it produces iterations that converge to a critical point on the generalized Stiefel manifold defined in expectation. The method has lower per-iteration cost, requires only matrix multiplications, and has the same convergence rates as its Riemannian counterparts involving the full matrix $B$. Experiments demonstrate its effectiveness in various machine learning applications involving generalized orthogonality constraints, including CCA, ICA, and GEVP.
Authors: Xavier Mart\'inez Lua\~na, Rebeca P. D\'iaz Redondo, Manuel Fern\'andez Veiga
Abstract: Federated Learning (FL) is an interesting strategy that enables the collaborative training of an AI model among different data owners without revealing their private datasets. Even so, FL has some privacy vulnerabilities that have been tried to be overcome by applying some techniques like Differential Privacy (DP), Homomorphic Encryption (HE), or Secure Multi-Party Computation (SMPC). However, these techniques have some important drawbacks that might narrow their range of application: problems to work with non-linear functions and to operate large matrix multiplications and high communication and computational costs to manage semi-honest nodes. In this context, we propose a solution to guarantee privacy in FL schemes that simultaneously solves the previously mentioned problems. Our proposal is based on the Berrut Approximated Coded Computing, a technique from the Coded Distributed Computing paradigm, adapted to a Secret Sharing configuration, to provide input privacy to FL in a scalable way. It can be applied for computing non-linear functions and treats the special case of distributed matrix multiplication, a key primitive at the core of many automated learning tasks. Because of these characteristics, it could be applied in a wide range of FL scenarios, since it is independent of the machine learning models or aggregation algorithms used in the FL scheme. We provide analysis of the achieve privacy and complexity of our solution and, due to the extensive numerical results performed, it can be observed a good trade-off between privacy and precision.
Authors: Kimia Kamal, Bilal Farooq
Abstract: Transport policy assessment often involves causal questions, yet the causal inference capabilities of traditional travel behavioural models are at best limited. We present the deep CAusal infeRence mOdel for traveL behavIour aNAlysis (CAROLINA), a framework that explicitly models causality in travel behaviour, enhances predictive accuracy, and maintains interpretability by leveraging causal inference, deep learning, and traditional discrete choice modelling. Within this framework, we introduce a Generative Counterfactual model for forecasting human behaviour by adapting the Normalizing Flow method. Through the case studies of virtual reality-based pedestrian crossing behaviour, revealed preference travel behaviour from London, and synthetic data, we demonstrate the effectiveness of our proposed models in uncovering causal relationships, prediction accuracy, and assessing policy interventions. Our results show that intervention mechanisms that can reduce pedestrian stress levels lead to a 38.5% increase in individuals experiencing shorter waiting times. Reducing the travel distances in London results in a 47% increase in sustainable travel modes.
Authors: Abdoul Jalil Djiberou Mahamadou, Lea Goetz, Russ Altman
Abstract: Inherent bias within society can be amplified and perpetuated by artificial intelligence (AI) systems. To address this issue, a wide range of solutions have been proposed to identify and mitigate bias and enforce fairness for individuals and groups. Recently, Graph Laplacian Regularizer (GLR), a regularization technique from the semi-supervised learning literature has been used as a substitute for the common Lipschitz condition to enhance individual fairness (IF). Notable prior work has shown that enforcing IF through a GLR can improve the transfer learning accuracy of AI models under covariate shifts. However, the prior work defines a GLR on the source and target data combined, implicitly assuming that the target data are available at train time, which might not hold in practice. In this work, we investigated whether defining a GLR independently on the train and target data could maintain similar accuracy compared to the prior work model. Furthermore, we introduced the Normalized Fairness Gain score (FGN) to measure IF for in-processing algorithmic fairness techniques. FGN quantifies the amount of gained fairness when a GLR is used versus not. We evaluated the new and original methods under FGN, the Prediction Consistency (PC), and traditional classification metrics on the German Credit Approval dataset. The results showed that the two models achieved similar statistical mean performances over five-fold cross-validation. Furthermore, the proposed metric showed that PC scores can be misleading as the scores can be high and statistically similar to fairness-enhanced models while FGN scores are small. This work therefore provides new insights into when a GLR effectively enhances IF and the pitfalls of PC.
Authors: Yuwei Liu, Chen Dan, Anubhav Bhatti, Bingjie Shen, Divij Gupta, Suraj Parmar, San Lee
Abstract: Sepsis is a leading cause of mortality in intensive care units (ICUs), representing a substantial medical challenge. The complexity of analyzing diverse vital signs to predict sepsis further aggravates this issue. While deep learning techniques have been advanced for early sepsis prediction, their 'black-box' nature obscures the internal logic, impairing interpretability in critical settings like ICUs. This paper introduces a framework that combines a deep learning model with an attention mechanism that highlights the critical time steps in the forecasting process, thus improving model interpretability and supporting clinical decision-making. We show that the attention mechanism could be adapted to various black box time series forecasting models such as N-HiTS and N-BEATS. Our method preserves the accuracy of conventional deep learning models while enhancing interpretability through attention-weight-generated heatmaps. We evaluated our model on the eICU-CRD dataset, focusing on forecasting vital signs for sepsis patients. We assessed its performance using mean squared error (MSE) and dynamic time warping (DTW) metrics. We explored the attention maps of N-HiTS and N-BEATS, examining the differences in their performance and identifying crucial factors influencing vital sign forecasting.
Authors: Xinyi Ni, Lifeng Lai
Abstract: Robust Markov Decision Processes (RMDPs) have received significant research interest, offering an alternative to standard Markov Decision Processes (MDPs) that often assume fixed transition probabilities. RMDPs address this by optimizing for the worst-case scenarios within ambiguity sets. While earlier studies on RMDPs have largely centered on risk-neutral reinforcement learning (RL), with the goal of minimizing expected total discounted costs, in this paper, we analyze the robustness of CVaR-based risk-sensitive RL under RMDP. Firstly, we consider predetermined ambiguity sets. Based on the coherency of CVaR, we establish a connection between robustness and risk sensitivity, thus, techniques in risk-sensitive RL can be adopted to solve the proposed problem. Furthermore, motivated by the existence of decision-dependent uncertainty in real-world problems, we study problems with state-action-dependent ambiguity sets. To solve this, we define a new risk measure named NCVaR and build the equivalence of NCVaR optimization and robust CVaR optimization. We further propose value iteration algorithms and validate our approach in simulation experiments.
Authors: Guanhua Zhang, Moritz Hardt
Abstract: We examine multi-task benchmarks in machine learning through the lens of social choice theory. We draw an analogy between benchmarks and electoral systems, where models are candidates and tasks are voters. This suggests a distinction between cardinal and ordinal benchmark systems. The former aggregate numerical scores into one model ranking; the latter aggregate rankings for each task. We apply Arrow's impossibility theorem to ordinal benchmarks to highlight the inherent limitations of ordinal systems, particularly their sensitivity to the inclusion of irrelevant models. Inspired by Arrow's theorem, we empirically demonstrate a strong trade-off between diversity and sensitivity to irrelevant changes in existing multi-task benchmarks. Our result is based on new quantitative measures of diversity and sensitivity that we introduce. Sensitivity quantifies the impact that irrelevant changes to tasks have on a benchmark. Diversity captures the degree of disagreement in model rankings across tasks. We develop efficient approximation algorithms for both measures, as exact computation is computationally challenging. Through extensive experiments on seven cardinal benchmarks and eleven ordinal benchmarks, we demonstrate a clear trade-off between diversity and stability: The more diverse a multi-task benchmark, the more sensitive to trivial changes it is. Additionally, we show that the aggregated rankings of existing benchmarks are highly unstable under irrelevant changes. The codes and data are available at https://socialfoundations.github.io/benchbench/.
Authors: Sam Reifenstein, Timothee Leleu, Yoshihisa Yamamoto
Abstract: We propose a novel algorithm that extends the methods of ball smoothing and Gaussian smoothing for noisy derivative-free optimization by accounting for the heterogeneous curvature of the objective function. The algorithm dynamically adapts the shape of the smoothing kernel to approximate the Hessian of the objective function around a local optimum. This approach significantly reduces the error in estimating the gradient from noisy evaluations through sampling. We demonstrate the efficacy of our method through numerical experiments on artificial problems. Additionally, we show improved performance when tuning NP-hard combinatorial optimization solvers compared to existing state-of-the-art heuristic derivative-free and Bayesian optimization methods.
Authors: Haiguang Li, Usama Pervaiz, Joseph Antognini, Micha{\l} Matuszak, Lawrence Au, Gilles Roux, Trausti Thormundsso
Abstract: On-device machine learning (ODML) enables powerful edge applications, but power consumption remains a key challenge for resource-constrained devices. To address this, developers often face a trade-off between model accuracy and power consumption, employing either computationally intensive models on high-power cores or pared-down models on low-power cores. Both approaches typically lead to a compromise in user experience (UX). This work focuses on the use of Gated Compression (GC) layer to enhance ODML model performance while conserving power and maximizing cost-efficiency, especially for always-on use cases. GC layers dynamically regulate data flow by selectively gating activations of neurons within the neural network and effectively filtering out non-essential inputs, which reduces power needs without compromising accuracy, and enables more efficient execution on heterogeneous compute cores. These improvements enhance UX through prolonged battery life, improved device responsiveness, and greater user comfort. In this work, we have integrated GC layers into vision and speech domain models including the transformer-based ViT model. Our experiments demonstrate theoretical power efficiency gains ranging from 158x to 30,000x for always-on scenarios. This substantial improvement empowers ODML applications with enhanced UX benefits.
Authors: Elahe Khatibi, Mahyar Abbasian, Zhongqi Yang, Iman Azimi, Amir M. Rahmani
Abstract: To perform effective causal inference in high-dimensional datasets, initiating the process with causal discovery is imperative, wherein a causal graph is generated based on observational data. However, obtaining a complete and accurate causal graph poses a formidable challenge, recognized as an NP-hard problem. Recently, the advent of Large Language Models (LLMs) has ushered in a new era, indicating their emergent capabilities and widespread applicability in facilitating causal reasoning across diverse domains, such as medicine, finance, and science. The expansive knowledge base of LLMs holds the potential to elevate the field of causal reasoning by offering interpretability, making inferences, generalizability, and uncovering novel causal structures. In this paper, we introduce a new framework, named Autonomous LLM-Augmented Causal Discovery Framework (ALCM), to synergize data-driven causal discovery algorithms and LLMs, automating the generation of a more resilient, accurate, and explicable causal graph. The ALCM consists of three integral components: causal structure learning, causal wrapper, and LLM-driven causal refiner. These components autonomously collaborate within a dynamic environment to address causal discovery questions and deliver plausible causal graphs. We evaluate the ALCM framework by implementing two demonstrations on seven well-known datasets. Experimental results demonstrate that ALCM outperforms existing LLM methods and conventional data-driven causal reasoning mechanisms. This study not only shows the effectiveness of the ALCM but also underscores new research directions in leveraging the causal reasoning capabilities of LLMs.
Authors: Marzi Heidari, Hanping Zhang, Yuhong Guo
Abstract: In recent years, semi-supervised learning (SSL) has gained significant attention due to its ability to leverage both labeled and unlabeled data to improve model performance, especially when labeled data is scarce. However, most current SSL methods rely on heuristics or predefined rules for generating pseudo-labels and leveraging unlabeled data. They are limited to exploiting loss functions and regularization methods within the standard norm. In this paper, we propose a novel Reinforcement Learning (RL) Guided SSL method, RLGSSL, that formulates SSL as a one-armed bandit problem and deploys an innovative RL loss based on weighted reward to adaptively guide the learning process of the prediction model. RLGSSL incorporates a carefully designed reward function that balances the use of labeled and unlabeled data to enhance generalization performance. A semi-supervised teacher-student framework is further deployed to increase the learning stability. We demonstrate the effectiveness of RLGSSL through extensive experiments on several benchmark datasets and show that our approach achieves consistent superior performance compared to state-of-the-art SSL methods.
Authors: Shengyao Lu, Bang Liu, Keith G. Mills, Jiao He, Di Niu
Abstract: Understanding and explaining the predictions of Graph Neural Networks (GNNs), is crucial for enhancing their safety and trustworthiness. Subgraph-level explanations are gaining attention for their intuitive appeal. However, most existing subgraph-level explainers face efficiency challenges in explaining GNNs due to complex search processes. The key challenge is to find a balance between intuitiveness and efficiency while ensuring transparency. Additionally, these explainers usually induce subgraphs by nodes, which may introduce less-intuitive disconnected nodes in the subgraph-level explanations or omit many important subgraph structures. In this paper, we reveal that inducing subgraph explanations by edges is more comprehensive than other subgraph inducing techniques. We also emphasize the need of determining the subgraph explanation size for each data instance, as different data instances may involve different important substructures. Building upon these considerations, we introduce a training-free approach, named EiG-Search. We employ an efficient linear-time search algorithm over the edge-induced subgraphs, where the edges are ranked by an enhanced gradient-based importance. We conduct extensive experiments on a total of seven datasets, demonstrating its superior performance and efficiency both quantitatively and qualitatively over the leading baselines.
Authors: Elvis Togban, Djemel Ziou
Abstract: This paper presents a discriminative classifier for compositional data. This classifier is based on the posterior distribution of the Generalized Dirichlet which is the discriminative counterpart of Generalized Dirichlet mixture model. Moreover, following the mixture of experts paradigm, we proposed a hierarchical mixture of this classifier. In order to learn the models parameters, we use a variational approximation by deriving an upper-bound for the Generalized Dirichlet mixture. To the best of our knownledge, this is the first time this bound is proposed in the literature. Experimental results are presented for spam detection and color space identification.
Authors: Shaoyuan Chen, Yutong Lin, Mingxing Zhang, Yongwei Wu
Abstract: Transformer-based large language models (LLMs) exhibit impressive performance in generative tasks but introduce significant challenges in real-world serving due to inefficient use of the expensive, computation-optimized accelerators. This mismatch arises from the autoregressive nature of LLMs, where the generation phase comprises operators with varying resource demands. Specifically, the attention operator is memory-intensive, exhibiting a memory access pattern that clashes with the strengths of modern accelerators, especially as context length increases. To enhance the efficiency and cost-effectiveness of LLM serving, we introduce the concept of attention offloading. This approach leverages a collection of cheap, memory-optimized devices for the attention operator while still utilizing high-end accelerators for other parts of the model. This heterogeneous setup ensures that each component is tailored to its specific workload, maximizing overall performance and cost efficiency. Our comprehensive analysis and experiments confirm the viability of splitting the attention computation over multiple devices. Also, the communication bandwidth required between heterogeneous devices proves to be manageable with prevalent networking technologies. To further validate our theory, we develop Lamina, an LLM inference system that incorporates attention offloading. Experimental results indicate that Lamina can provide 1.48x-12.1x higher estimated throughput per dollar than homogeneous solutions.
Authors: Jiancong Xiao, Jiawei Zhang, Zhi-Quan Luo, Asuman Ozdaglar
Abstract: In adversarial machine learning, neural networks suffer from a significant issue known as robust overfitting, where the robust test accuracy decreases over epochs (Rice et al., 2020). Recent research conducted by Xing et al.,2021; Xiao et al., 2022 has focused on studying the uniform stability of adversarial training. Their investigations revealed that SGD-based adversarial training fails to exhibit uniform stability, and the derived stability bounds align with the observed phenomenon of robust overfitting in experiments. This motivates us to develop uniformly stable algorithms specifically tailored for adversarial training. To this aim, we introduce Moreau envelope-$\mathcal{A}$, a variant of the Moreau Envelope-type algorithm. We employ a Moreau envelope function to reframe the original problem as a min-min problem, separating the non-strong convexity and non-smoothness of the adversarial loss. Then, this approach alternates between solving the inner and outer minimization problems to achieve uniform stability without incurring additional computational overhead. In practical scenarios, we show the efficacy of ME-$\mathcal{A}$ in mitigating the issue of robust overfitting. Beyond its application in adversarial training, this represents a fundamental result in uniform stability analysis, as ME-$\mathcal{A}$ is the first algorithm to exhibit uniform stability for weakly-convex, non-smooth problems.
Authors: Trinath Sai Subhash Reddy Pittala, Uma Maheswara Rao Meleti, Geethakrishna Puligundla
Abstract: Recent developments in adversarial machine learning have highlighted the importance of building robust AI systems to protect against increasingly sophisticated attacks. While frameworks like AI Guardian are designed to defend against these threats, they often rely on assumptions that can limit their effectiveness. For example, they may assume attacks only come from one direction or include adversarial images in their training data. Our proposal suggests a different approach to the AI Guardian framework. Instead of including adversarial examples in the training process, we propose training the AI system without them. This aims to create a system that is inherently resilient to a wider range of attacks. Our method focuses on a dynamic defense strategy using stable diffusion that learns continuously and models threats comprehensively. We believe this approach can lead to a more generalized and robust defense against adversarial attacks. In this paper, we outline our proposed approach, including the theoretical basis, experimental design, and expected impact on improving AI security against adversarial threats.
Authors: Mudit Gaur, Vaneet Aggarwal, Amrit Singh Bedi, Di Wang
Abstract: The current state-of-the-art theoretical analysis of Actor-Critic (AC) algorithms significantly lags in addressing the practical aspects of AC implementations. This crucial gap needs bridging to bring the analysis in line with practical implementations of AC. To address this, we advocate for considering the MMCLG criteria: \textbf{M}ulti-layer neural network parametrization for actor/critic, \textbf{M}arkovian sampling, \textbf{C}ontinuous state-action spaces, the performance of the \textbf{L}ast iterate, and \textbf{G}lobal optimality. These aspects are practically significant and have been largely overlooked in existing theoretical analyses of AC algorithms. In this work, we address these gaps by providing the first comprehensive theoretical analysis of AC algorithms that encompasses all five crucial practical aspects (covers MMCLG criteria). We establish global convergence sample complexity bounds of $\tilde{\mathcal{O}}\left({\epsilon^{-3}}\right)$. We achieve this result through our novel use of the weak gradient domination property of MDP's and our unique analysis of the error in critic estimation.
Authors: Sicong Liu, Wentao Zhou, Zimu Zhou, Bin Guo, Minfan Wang, Cheng Fang, Zheng Lin, Zhiwen Yu
Abstract: There is a growing demand to deploy computation-intensive deep learning (DL) models on resource-constrained mobile devices for real-time intelligent applications. Equipped with a variety of processing units such as CPUs, GPUs, and NPUs, the mobile devices hold potential to accelerate DL inference via parallel execution across heterogeneous processors. Various efficient parallel methods have been explored to optimize computation distribution, achieve load balance, and minimize communication cost across processors. Yet their practical effectiveness in the dynamic and diverse real-world mobile environment is less explored. This paper presents a holistic empirical study to assess the capabilities and challenges associated with parallel DL inference on heterogeneous mobile processors. Through carefully designed experiments covering various DL models, mobile software/hardware environments, workload patterns, and resource availability, we identify limitations of existing techniques and highlight opportunities for cross-level optimization.
Authors: Ziang Zhou, Jieming Shi, Renchi Yang, Yuanhang Zou, Qing Li
Abstract: Heterogeneous graphs are ubiquitous to model complex data. There are urgent needs on powerful heterogeneous graph neural networks to effectively support important applications. We identify a potential semantic mixing issue in existing message passing processes, where the representations of the neighbors of a node $v$ are forced to be transformed to the feature space of $v$ for aggregation, though the neighbors are in different types. That is, the semantics in different node types are entangled together into node $v$'s representation. To address the issue, we propose SlotGAT with separate message passing processes in slots, one for each node type, to maintain the representations in their own node-type feature spaces. Moreover, in a slot-based message passing layer, we design an attention mechanism for effective slot-wise message aggregation. Further, we develop a slot attention technique after the last layer of SlotGAT, to learn the importance of different slots in downstream tasks. Our analysis indicates that the slots in SlotGAT can preserve different semantics in various feature spaces. The superiority of SlotGAT is evaluated against 13 baselines on 6 datasets for node classification and link prediction. Our code is at https://github.com/scottjiao/SlotGAT_ICML23/.
Authors: Sung Moon Ko, Sumin Lee, Dae-Woong Jeong, Hyunseung Kim, Chanhui Lee, Soorin Yim, Sehui Han
Abstract: Molecular datasets often suffer from a lack of data. It is well-known that gathering data is difficult due to the complexity of experimentation or simulation involved. Here, we leverage mutual information across different tasks in molecular data to address this issue. We extend an algorithm that utilizes the geometric characteristics of the encoding space, known as the Geometrically Aligned Transfer Encoder (GATE), to a multi-task setup. Thus, we connect multiple molecular tasks by aligning the curved coordinates onto locally flat coordinates, ensuring the flow of information from source tasks to support performance on target data.
Authors: Vegard Flovik
Abstract: Distribution shifts, where statistical properties differ between training and test datasets, present a significant challenge in real-world machine learning applications where they directly impact model generalization and robustness. In this study, we explore model adaptation and generalization by utilizing synthetic data to systematically address distributional disparities. Our investigation aims to identify the prerequisites for successful model adaptation across diverse data distributions, while quantifying the associated uncertainties. Specifically, we generate synthetic data using the Van der Waals equation for gases and employ quantitative measures such as Kullback-Leibler divergence, Jensen-Shannon distance, and Mahalanobis distance to assess data similarity. These metrics en able us to evaluate both model accuracy and quantify the associated uncertainty in predictions arising from data distribution shifts. Our findings suggest that utilizing statistical measures, such as the Mahalanobis distance, to determine whether model predictions fall within the low-error "interpolation regime" or the high-error "extrapolation regime" provides a complementary method for assessing distribution shift and model uncertainty. These insights hold significant value for enhancing model robustness and generalization, essential for the successful deployment of machine learning applications in real-world scenarios.
Authors: Puning Zhao, Jintao Deng, Xu Cheng
Abstract: PU learning refers to the classification problem in which only part of positive samples are labeled. Existing PU learning methods treat unlabeled samples equally. However, in many real tasks, from common sense or domain knowledge, some unlabeled samples are more likely to be positive than others. In this paper, we propose soft label PU learning, in which unlabeled data are assigned soft labels according to their probabilities of being positive. Considering that the ground truth of TPR, FPR, and AUC are unknown, we then design PU counterparts of these metrics to evaluate the performances of soft label PU learning methods within validation data. We show that these new designed PU metrics are good substitutes for the real metrics. After that, a method that optimizes such metrics is proposed. Experiments on public datasets and real datasets for anti-cheat services from Tencent games demonstrate the effectiveness of our proposed method.
Authors: S. Savazzi, V. Rampa, S. Kianoush, A. Minora, L. Costa
Abstract: The paper considers the problem of human-scale RF sensing utilizing a network of resource-constrained MIMO radars with low range-azimuth resolution. The radars operate in the mmWave band and obtain time-varying 3D point cloud (PC) information that is sensitive to body movements. They also observe the same scene from different views and cooperate while sensing the environment using a sidelink communication channel. Conventional cooperation setups allow the radars to mutually exchange raw PC information to improve ego sensing. The paper proposes a federation mechanism where the radars exchange the parameters of a Bayesian posterior measure of the observed PCs, rather than raw data. The radars act as distributed parameter servers to reconstruct a global posterior (i.e., federated posterior) using Bayesian tools. The paper quantifies and compares the benefits of radar federation with respect to cooperation mechanisms. Both approaches are validated by experiments with a real-time demonstration platform. Federation makes minimal use of the sidelink communication channel (20 {\div} 25 times lower bandwidth use) and is less sensitive to unresolved targets. On the other hand, cooperation reduces the mean absolute target estimation error of about 20%.
Authors: Patrick Schnell, Nils Thuerey
Abstract: Of all the vector fields surrounding the minima of recurrent learning setups, the gradient field with its exploding and vanishing updates appears a poor choice for optimization, offering little beyond efficient computability. We seek to improve this suboptimal practice in the context of physics simulations, where backpropagating feedback through many unrolled time steps is considered crucial to acquiring temporally coherent behavior. The alternative vector field we propose follows from two principles: physics simulators, unlike neural networks, have a balanced gradient flow, and certain modifications to the backpropagation pass leave the positions of the original minima unchanged. As any modification of backpropagation decouples forward and backward pass, the rotation-free character of the gradient field is lost. Therefore, we discuss the negative implications of using such a rotational vector field for optimization and how to counteract them. Our final procedure is easily implementable via a sequence of gradient stopping and component-wise comparison operations, which do not negatively affect scalability. Our experiments on three control problems show that especially as we increase the complexity of each task, the unbalanced updates from the gradient can no longer provide the precise control signals necessary while our method still solves the tasks. Our code can be found at https://github.com/tum-pbs/StableBPTT.
Authors: Anton Plaksin, Vitaly Kalev
Abstract: Robust Reinforcement Learning (RRL) is a promising Reinforcement Learning (RL) paradigm aimed at training robust to uncertainty or disturbances models, making them more efficient for real-world applications. Following this paradigm, uncertainty or disturbances are interpreted as actions of a second adversarial agent, and thus, the problem is reduced to seeking the agents' policies robust to any opponent's actions. This paper is the first to propose considering the RRL problems within the positional differential game theory, which helps us to obtain theoretically justified intuition to develop a centralized Q-learning approach. Namely, we prove that under Isaacs's condition (sufficiently general for real-world dynamical systems), the same Q-function can be utilized as an approximate solution of both minimax and maximin Bellman equations. Based on these results, we present the Isaacs Deep Q-Network algorithms and demonstrate their superiority compared to other baseline RRL and Multi-Agent RL algorithms in various environments.
Authors: William Lindskog, Christian Prehofer
Abstract: In this paper, we show how Federated Learning (FL) can be applied to vehicular use-cases in which we seek to classify obstacles, irregularities and pavement types on roads. Our proposed framework utilizes FL and TabNet, a state-of-the-art neural network for tabular data. We are the first to demonstrate how TabNet can be integrated with FL. Moreover, we achieve a maximum test accuracy of 93.6%. Finally, we reason why FL is a suitable concept for this data set.
Authors: Yichuan Zou, Li Jin, Xi Xiong
Abstract: Platooning of connected and autonomous vehicles (CAVs) plays a vital role in modernizing highways, ushering in enhanced efficiency and safety. This paper explores the significance of platooning in smart highways, employing a coupled partial differential equation (PDE) and ordinary differential equation (ODE) model to elucidate the complex interaction between bulk traffic flow and CAV platoons. Our study focuses on developing a Dyna-style planning and learning framework tailored for platoon control, with a specific goal of reducing fuel consumption. By harnessing the coupled PDE-ODE model, we improve data efficiency in Dyna-style learning through virtual experiences. Simulation results validate the effectiveness of our macroscopic model in modeling platoons within mixed-autonomy settings, demonstrating a notable $10.11\%$ reduction in vehicular fuel consumption compared to conventional approaches.
Authors: David J. Schodt
Abstract: Bayesian Neural Networks (BNNs) extend traditional neural networks to provide uncertainties associated with their outputs. On the forward pass through a BNN, predictions (and their uncertainties) are made either by Monte Carlo sampling network weights from the learned posterior or by analytically propagating statistical moments through the network. Though flexible, Monte Carlo sampling is computationally expensive and can be infeasible or impractical under resource constraints or for large networks. While moment propagation can ameliorate the computational costs of BNN inference, it can be difficult or impossible for networks with arbitrary nonlinearities, thereby restricting the possible set of network layers permitted with such a scheme. In this work, we demonstrate a simple yet effective approach for propagating statistical moments through arbitrary nonlinearities with only 3 deterministic samples, enabling few-sample variational inference of BNNs without restricting the set of network layers used. Furthermore, we leverage this approach to demonstrate a novel nonlinear activation function that we use to inject physics-informed prior information into output nodes of a BNN.
Authors: William Lindskog, Christian Prehofer, Sarandeep Singh
Abstract: Federated Learning (FL) has gained considerable traction, yet, for tabular data, FL has received less attention. Most FL research has focused on Neural Networks while Tree-Based Models (TBMs) such as XGBoost have historically performed better on tabular data. It has been shown that subsampling of training data when building trees can improve performance but it is an open problem whether such subsampling can improve performance in FL. In this paper, we evaluate a histogram-based federated XGBoost that uses Minimal Variance Sampling (MVS). We demonstrate the underlying algorithm and show that our model using MVS can improve performance in terms of accuracy and regression error in a federated setting. In our evaluation, our model using MVS performs better than uniform (random) sampling and no sampling at all. It achieves both outstanding local and global performance on a new set of federated tabular datasets. Federated XGBoost using MVS also outperforms centralized XGBoost in half of the studied cases.
Authors: William Lindskog, Christian Prehofer
Abstract: Federated Learning (FL) has lately gained traction as it addresses how machine learning models train on distributed datasets. FL was designed for parametric models, namely Deep Neural Networks (DNNs).Thus, it has shown promise on image and text tasks. However, FL for tabular data has received little attention. Tree-Based Models (TBMs) have been considered to perform better on tabular data and they are starting to see FL integrations. In this study, we benchmark federated TBMs and DNNs for horizontal FL, with varying data partitions, on 10 well-known tabular datasets. Our novel benchmark results indicates that current federated boosted TBMs perform better than federated DNNs in different data partitions. Furthermore, a federated XGBoost outperforms all other models. Lastly, we find that federated TBMs perform better than federated parametric models, even when increasing the number of clients significantly.
Authors: Christos Louizos, Matthias Reisser, Denis Korzhenkov
Abstract: We investigate contrastive learning in the federated setting through the lens of SimCLR and multi-view mutual information maximization. In doing so, we uncover a connection between contrastive representation learning and user verification; by adding a user verification loss to each client's local SimCLR loss we recover a lower bound to the global multi-view mutual information. To accommodate for the case of when some labelled data are available at the clients, we extend our SimCLR variant to the federated semi-supervised setting. We see that a supervised SimCLR objective can be obtained with two changes: a) the contrastive loss is computed between datapoints that share the same label and b) we require an additional auxiliary head that predicts the correct labels from either of the two views. Along with the proposed SimCLR extensions, we also study how different sources of non-i.i.d.-ness can impact the performance of federated unsupervised learning through global mutual information maximization; we find that a global objective is beneficial for some sources of non-i.i.d.-ness but can be detrimental for others. We empirically evaluate our proposed extensions in various tasks to validate our claims and furthermore demonstrate that our proposed modifications generalize to other pretraining methods.
Authors: Guillaume Perez, Michel Barlaud
Abstract: The $\ell_{1,\infty}$ norm is an efficient structured projection but the complexity of the best algorithm is unfortunately $\mathcal{O}\big(n m \log(n m)\big)$ for a matrix in $\mathbb{R}^{n\times m}$. In this paper, we propose a new bi-level projection method for which we show that the time complexity for the $\ell_{1,\infty}$ norm is only $\mathcal{O}\big(n m \big)$ for a matrix in $\mathbb{R}^{n\times m}$, and $\mathcal{O}\big(n + m \big)$ with full parallel power. We generalize our method to tensors and we propose a new multi-level projection, having an induced decomposition that yields a linear parallel speedup up to an exponential speedup factor, resulting in a time complexity lower-bounded by the sum of the dimensions. Experiments show that our bi-level $\ell_{1,\infty}$ projection is $2.5$ times faster than the actual fastest algorithm provided by \textit{Chu et. al.} while providing same accuracy and better sparsity in neural networks applications.
Authors: Daniel Fesalbon
Abstract: With recent studies related to Neural Networks being used on different forecasting and time series investigations, this study aims to expand these contexts to ferry passenger traffic. The primary objective of the study is to investigate and evaluate an LSTM-based Neural Networks' capability to forecast ferry passengers of two ports in the Philippines. The proposed model's fitting and evaluation of the passenger flow forecasting of the two ports is based on monthly passenger traffic from 2016 to 2022 data that was acquired from the Philippine Ports Authority (PPA). This work uses Mean Absolute Percentage Error (MAPE) as its primary metric to evaluate the model's forecasting capability. The proposed LSTM-based Neural Networks model achieved 72% forecasting accuracy to the Batangas port ferry passenger data and 74% forecasting accuracy to the Mindoro port ferry passenger data. Using Keras and Scikit-learn Python libraries, this work concludes a reasonable forecasting performance of the presented LSTM model. Aside from these notable findings, this study also recommends further investigation and studies on employing other statistical, machine learning, and deep learning methods on forecasting ferry passenger flows.
Authors: Alvaro H. C. Correia, Fabio Valerio Massoli, Christos Louizos, Arash Behboodi
Abstract: Conformal Prediction (CP) is a distribution-free uncertainty estimation framework that constructs prediction sets guaranteed to contain the true answer with a user-specified probability. Intuitively, the size of the prediction set encodes a general notion of uncertainty, with larger sets associated with higher degrees of uncertainty. In this work, we leverage information theory to connect conformal prediction to other notions of uncertainty. More precisely, we prove three different ways to upper bound the intrinsic uncertainty, as described by the conditional entropy of the target variable given the inputs, by combining CP with information theoretical inequalities. Moreover, we demonstrate two direct and useful applications of such connection between conformal prediction and information theory: (i) more principled and effective conformal training objectives that generalize previous approaches and enable end-to-end training of machine learning models from scratch, and (ii) a natural mechanism to incorporate side information into conformal prediction. We empirically validate both applications in centralized and federated learning settings, showing our theoretical results translate to lower inefficiency (average prediction set size) for popular CP methods.
Authors: Roussel Desmond Nzoyem, David A. W. Barton, Tom Deakin
Abstract: Neural Ordinary Differential Equations typically struggle to generalize to new dynamical behaviors created by parameter changes in the underlying system, even when the dynamics are close to previously seen behaviors. The issue gets worse when the changing parameters are unobserved, i.e., their value or influence is not directly measurable when collecting data. We introduce Neural Context Flow (NCF), a framework that encodes said unobserved parameters in a latent context vector as input to a vector field. NCFs leverage differentiability of the vector field with respect to the parameters, along with first-order Taylor expansion to allow any context vector to influence trajectories from other parameters. We validate our method and compare it to established Multi-Task and Meta-Learning alternatives, showing competitive performance in mean squared error for in-domain and out-of-distribution evaluation on the Lotka-Volterra, Glycolytic Oscillator, and Gray-Scott problems. This study holds practical implications for foundational models in science and related areas that benefit from conditional neural ODEs. Our code is openly available at https://github.com/ddrous/ncflow.
Authors: Simone Brusatin, Tommaso Padoan, Andrea Coletta, Domenico Delli Gatti, Aldo Glielmo
Abstract: Agent-based models (ABMs) are simulation models used in economics to overcome some of the limitations of traditional frameworks based on general equilibrium assumptions. However, agents within an ABM follow predetermined, not fully rational, behavioural rules which can be cumbersome to design and difficult to justify. Here we leverage multi-agent reinforcement learning (RL) to expand the capabilities of ABMs with the introduction of fully rational agents that learn their policy by interacting with the environment and maximising a reward function. Specifically, we propose a 'Rational macro ABM' (R-MABM) framework by extending a paradigmatic macro ABM from the economic literature. We show that gradually substituting ABM firms in the model with RL agents, trained to maximise profits, allows for a thorough study of the impact of rationality on the economy. We find that RL agents spontaneously learn three distinct strategies for maximising profits, with the optimal strategy depending on the level of market competition and rationality. We also find that RL agents with independent policies, and without the ability to communicate with each other, spontaneously learn to segregate into different strategic groups, thus increasing market power and overall profits. Finally, we find that a higher degree of rationality in the economy always improves the macroeconomic environment as measured by total output, depending on the specific rational policy, this can come at the cost of higher instability. Our R-MABM framework is general, it allows for stable multi-agent learning, and represents a principled and robust direction to extend existing economic simulators.
Authors: Weijie Xia, Chenguang Wang, Peter Palensky, Pedro P. Vergara
Abstract: Residential Load Profile (RLP) generation and prediction are critical for the operation and planning of distribution networks, particularly as diverse low-carbon technologies are increasingly integrated. This paper introduces a novel flow-based generative model, termed Full Convolutional Profile Flow (FCPFlow), which is uniquely designed for both conditional and unconditional RLP generation, and for probabilistic load forecasting. By introducing two new layers--the invertible linear layer and the invertible normalization layer--the proposed FCPFlow architecture shows three main advantages compared to traditional statistical and contemporary deep generative models: 1) it is well-suited for RLP generation under continuous conditions, such as varying weather and annual electricity consumption, 2) it shows superior scalability in different datasets compared to traditional statistical, and 3) it also demonstrates better modeling capabilities in capturing the complex correlation of RLPs compared with deep generative models.
Authors: Luca Viano, Stratis Skoulakis, Volkan Cevher
Abstract: We present a new algorithm for imitation learning in infinite horizon linear MDPs dubbed ILARL which greatly improves the bound on the number of trajectories that the learner needs to sample from the environment. In particular, we remove exploration assumptions required in previous works and we improve the dependence on the desired accuracy $\epsilon$ from $\mathcal{O}\br{\epsilon^{-5}}$ to $\mathcal{O}\br{\epsilon^{-4}}$. Our result relies on a connection between imitation learning and online learning in MDPs with adversarial losses. For the latter setting, we present the first result for infinite horizon linear MDP which may be of independent interest. Moreover, we are able to provide a strengthen result for the finite horizon case where we achieve $\mathcal{O}\br{\epsilon^{-2}}$. Numerical experiments with linear function approximation shows that ILARL outperforms other commonly used algorithms.
Authors: Toon Vanderschueren, Wouter Verbeke, Felipe Moraes, Hugo Manuel Proen\c{c}a
Abstract: Efficiently allocating treatments with a budget constraint constitutes an important challenge across various domains. In marketing, for example, the use of promotions to target potential customers and boost conversions is limited by the available budget. While much research focuses on estimating causal effects, there is relatively limited work on learning to allocate treatments while considering the operational context. Existing methods for uplift modeling or causal inference primarily estimate treatment effects, without considering how this relates to a profit maximizing allocation policy that respects budget constraints. The potential downside of using these methods is that the resulting predictive model is not aligned with the operational context. Therefore, prediction errors are propagated to the optimization of the budget allocation problem, subsequently leading to a suboptimal allocation policy. We propose an alternative approach based on learning to rank. Our proposed methodology directly learns an allocation policy by prioritizing instances in terms of their incremental profit. We propose an efficient sampling procedure for the optimization of the ranking model to scale our methodology to large-scale data sets. Theoretically, we show how learning to rank can maximize the area under a policy's incremental profit curve. Empirically, we validate our methodology and show its effectiveness in practice through a series of experiments on both synthetic and real-world data.
Authors: Moritz Herrmann, F. Julian D. Lange, Katharina Eggensperger, Giuseppe Casalicchio, Marcel Wever, Matthias Feurer, David R\"ugamer, Eyke H\"ullermeier, Anne-Laure Boulesteix, Bernd Bischl
Abstract: We warn against a common but incomplete understanding of empirical research in machine learning (ML) that leads to non-replicable results, makes findings unreliable, and threatens to undermine progress in the field. To overcome this alarming situation, we call for more awareness of the plurality of ways of gaining knowledge experimentally but also of some epistemic limitations. In particular, we argue most current empirical ML research is fashioned as confirmatory research while it should rather be considered exploratory.
Authors: Alessandro Montenegro, Marco Mussi, Alberto Maria Metelli, Matteo Papini
Abstract: Policy gradient (PG) methods are successful approaches to deal with continuous reinforcement learning (RL) problems. They learn stochastic parametric (hyper)policies by either exploring in the space of actions or in the space of parameters. Stochastic controllers, however, are often undesirable from a practical perspective because of their lack of robustness, safety, and traceability. In common practice, stochastic (hyper)policies are learned only to deploy their deterministic version. In this paper, we make a step towards the theoretical understanding of this practice. After introducing a novel framework for modeling this scenario, we study the global convergence to the best deterministic policy, under (weak) gradient domination assumptions. Then, we illustrate how to tune the exploration level used for learning to optimize the trade-off between the sample complexity and the performance of the deployed deterministic policy. Finally, we quantitatively compare action-based and parameter-based exploration, giving a formal guise to intuitive results.
Authors: Elika Bozorgi, Saber Soleimani, Sakher Khalil Alqaiidi, Hamid Reza Arabnia, Krzysztof Kochut
Abstract: Graph is an important data representation which occurs naturally in the real world applications \cite{goyal2018graph}. Therefore, analyzing graphs provides users with better insights in different areas such as anomaly detection \cite{ma2021comprehensive}, decision making \cite{fan2023graph}, clustering \cite{tsitsulin2023graph}, classification \cite{wang2021mixup} and etc. However, most of these methods require high levels of computational time and space. We can use other ways like embedding to reduce these costs. Knowledge graph (KG) embedding is a technique that aims to achieve the vector representation of a KG. It represents entities and relations of a KG in a low-dimensional space while maintaining the semantic meanings of them. There are different methods for embedding graphs including random walk-based methods such as node2vec, metapath2vec and regpattern2vec. However, most of these methods bias the walks based on a rigid pattern usually hard-coded in the algorithm. In this work, we introduce \textit{subgraph2vec} for embedding KGs where walks are run inside a user-defined subgraph. We use this embedding for link prediction and prove our method has better performance in most cases in comparison with the previous ones.
Authors: Aaron Klein, Jacek Golebiowski, Xingchen Ma, Valerio Perrone, Cedric Archambeau
Abstract: Pre-trained language models (PLM), for example BERT or RoBERTa, mark the state-of-the-art for natural language understanding task when fine-tuned on labeled data. However, their large size poses challenges in deploying them for inference in real-world applications, due to significant GPU memory requirements and high inference latency. This paper explores neural architecture search (NAS) for structural pruning to find sub-parts of the fine-tuned network that optimally trade-off efficiency, for example in terms of model size or latency, and generalization performance. We also show how we can utilize more recently developed two-stage weight-sharing NAS approaches in this setting to accelerate the search process. Unlike traditional pruning methods with fixed thresholds, we propose to adopt a multi-objective approach that identifies the Pareto optimal set of sub-networks, allowing for a more flexible and automated compression process.
Authors: Sridhar Mahadevan
Abstract: Alan Turing proposed in 1950 a framework called an imitation game to decide if a machine could think. Using mathematics developed largely after Turing -- category theory -- we analyze a broader class of universal imitation games (UIGs), which includes static, dynamic, and evolutionary games. In static games, the participants are in a steady state. In dynamic UIGs, "learner" participants are trying to imitate "teacher" participants over the long run. In evolutionary UIGs, the participants are competing against each other in an evolutionary game, and participants can go extinct and be replaced by others with higher fitness. We use the framework of category theory -- in particular, two influential results by Yoneda -- to characterize each type of imitation game. Universal properties in categories are defined by initial and final objects. We characterize dynamic UIGs where participants are learning by inductive inference as initial algebras over well-founded sets, and contrast them with participants learning by conductive inference over the final coalgebra of non-well-founded sets. We briefly discuss the extension of our categorical framework for UIGs to imitation games on quantum computers.
Authors: Yicheng Zhan, Liang Shi, Wojciech Matusik, Qi Sun, Kaan Ak\c{s}it
Abstract: In the pursuit of advancing holographic display technology, we face a unique yet persistent roadblock: the inflexibility of learned holography in adapting to various hardware configurations. This is due to the variances in the complex optical components and system settings in existing holographic displays. Although the emerging learned approaches have enabled rapid and high-quality hologram generation, any alteration in display hardware still requires a retraining of the model. Our work introduces a configurable learned model that interactively computes 3D holograms from RGB-only 2D images for a variety of holographic displays. The model can be conditioned to predefined hardware parameters of existing holographic displays such as working wavelengths, pixel pitch, propagation distance, and peak brightness without having to retrain. In addition, our model accommodates various hologram types, including conventional single-color and emerging multi-color holograms that simultaneously use multiple color primaries in holographic displays. Notably, we enabled our hologram computations to rely on identifying the correlation between depth estimation and 3D hologram synthesis tasks within the learning domain for the first time in the literature. We employ knowledge distillation via a student-teacher learning strategy to streamline our model for interactive performance. Achieving up to a 2x speed improvement compared to state-of-the-art models while consistently generating high-quality 3D holograms with different hardware configurations.
Authors: Konstantin Grotov, Sergey Titov, Yaroslav Zharov, Timofey Bryksin
Abstract: Computational notebooks became indispensable tools for research-related development, offering unprecedented interactivity and flexibility in the development process. However, these benefits come at the cost of reproducibility and an increased potential for bugs. There are many tools for bug fixing; however, they are generally targeted at the classical linear code. With the rise of code-fluent Large Language Models, a new stream of smart bug-fixing tools has emerged. However, the applicability of those tools is still problematic for non-linear computational notebooks. In this paper, we propose a potential solution for resolving errors in computational notebooks via an iterative LLM-based agent. We discuss the questions raised by this approach and share a novel dataset of computational notebooks containing bugs to facilitate the research of the proposed approach.
Authors: Olli J\"arviniemi, Evan Hubinger
Abstract: We study the tendency of AI systems to deceive by constructing a realistic simulation setting of a company AI assistant. The simulated company employees provide tasks for the assistant to complete, these tasks spanning writing assistance, information retrieval and programming. We then introduce situations where the model might be inclined to behave deceptively, while taking care to not instruct or otherwise pressure the model to do so. Across different scenarios, we find that Claude 3 Opus 1) complies with a task of mass-generating comments to influence public perception of the company, later deceiving humans about it having done so, 2) lies to auditors when asked questions, and 3) strategically pretends to be less capable than it is during capability evaluations. Our work demonstrates that even models trained to be helpful, harmless and honest sometimes behave deceptively in realistic scenarios, without notable external pressure to do so.
Authors: Tanmay Sen, Ansuman Das, Mrinmay Sen
Abstract: Hate speech encompasses verbal, written, or behavioral communication that targets derogatory or discriminatory language against individuals or groups based on sensitive characteristics. Automated hate speech detection plays a crucial role in curbing its propagation, especially across social media platforms. Various methods, including recent advancements in deep learning, have been devised to address this challenge. In this study, we introduce HateTinyLLM, a novel framework based on fine-tuned decoder-only tiny large language models (tinyLLMs) for efficient hate speech detection. Our experimental findings demonstrate that the fine-tuned HateTinyLLM outperforms the pretrained mixtral-7b model by a significant margin. We explored various tiny LLMs, including PY007/TinyLlama-1.1B-step-50K-105b, Microsoft/phi-2, and facebook/opt-1.3b, and fine-tuned them using LoRA and adapter methods. Our observations indicate that all LoRA-based fine-tuned models achieved over 80\% accuracy.
Authors: Charlotte Van Petegem, Kasper Demeyere, Rien Maertens, Niko Strijbol, Bram De Wever, Bart Mesuere, Peter Dawyndt
Abstract: In programming education, providing manual feedback is essential but labour-intensive, posing challenges in consistency and timeliness. We introduce ECHO, a machine learning method to automate the reuse of feedback in educational code reviews by analysing patterns in abstract syntax trees. This study investigates two primary questions: whether ECHO can predict feedback annotations to specific lines of student code based on previously added annotations by human reviewers (RQ1), and whether its training and prediction speeds are suitable for using ECHO for real-time feedback during live code reviews by human reviewers (RQ2). Our results, based on annotations from both automated linting tools and human reviewers, show that ECHO can accurately and quickly predict appropriate feedback annotations. Its efficiency in processing and its flexibility in adapting to feedback patterns can significantly reduce the time and effort required for manual feedback provisioning in educational settings.
Authors: Vasu Sharma, Karthik Padthe, Newsha Ardalani, Kushal Tirumala, Russell Howes, Hu Xu, Po-Yao Huang, Shang-Wen Li, Armen Aghajanyan, Gargi Ghosh
Abstract: In recent times training Language Models (LMs) have relied on computationally heavy training over massive datasets which makes this training process extremely laborious. In this paper we propose a novel method for numerically evaluating text quality in large unlabelled NLP datasets in a model agnostic manner to assign the text instances a "quality score". By proposing the text quality metric, the paper establishes a framework to identify and eliminate low-quality text instances, leading to improved training efficiency for LM models. Experimental results over multiple models and datasets demonstrate the efficacy of this approach, showcasing substantial gains in training effectiveness and highlighting the potential for resource-efficient LM training. For example, we observe an absolute accuracy improvement of 0.9% averaged over 14 downstream evaluation tasks for multiple LM models while using 40% lesser data and training 42% faster when training on the OpenWebText dataset and 0.8% average absolute accuracy improvement while using 20% lesser data and training 21% faster on the Wikipedia dataset.
Authors: Nadia Saeed
Abstract: The MEDIQA-M3G 2024 challenge necessitates novel solutions for Multilingual & Multimodal Medical Answer Generation in dermatology (wai Yim et al., 2024a). This paper addresses the limitations of traditional methods by proposing a weakly supervised learning approach for open-ended medical question-answering (QA). Our system leverages readily available MEDIQA-M3G images via a VGG16-CNN-SVM model, enabling multilingual (English, Chinese, Spanish) learning of informative skin condition representations. Using pre-trained QA models, we further bridge the gap between visual and textual information through multimodal fusion. This approach tackles complex, open-ended questions even without predefined answer choices. We empower the generation of comprehensive answers by feeding the ViT-CLIP model with multiple responses alongside images. This work advances medical QA research, paving the way for clinical decision support systems and ultimately improving healthcare delivery.
Authors: Li Wan, Tansu Alpcan, Margreta Kuijper, Emanuele Viterbo
Abstract: We propose a novel, lightweight supervised dictionary learning framework for text classification based on data compression and representation. This two-phase algorithm initially employs the Lempel-Ziv-Welch (LZW) algorithm to construct a dictionary from text datasets, focusing on the conceptual significance of dictionary elements. Subsequently, dictionaries are refined considering label data, optimizing dictionary atoms to enhance discriminative power based on mutual information and class distribution. This process generates discriminative numerical representations, facilitating the training of simple classifiers such as SVMs and neural networks. We evaluate our algorithm's information-theoretic performance using information bottleneck principles and introduce the information plane area rank (IPAR) as a novel metric to quantify the information-theoretic performance. Tested on six benchmark text datasets, our algorithm competes closely with top models, especially in limited-vocabulary contexts, using significantly fewer parameters. \review{Our algorithm closely matches top-performing models, deviating by only ~2\% on limited-vocabulary datasets, using just 10\% of their parameters. However, it falls short on diverse-vocabulary datasets, likely due to the LZW algorithm's constraints with low-repetition data. This contrast highlights its efficiency and limitations across different dataset types.
Authors: Nidhi Kamal, Saurabh Yadav, Jorawar Singh, Aditi Avasthi
Abstract: Providing fast and accurate resolution to the student's query is an essential solution provided by Edtech organizations. This is generally provided with a chat-bot like interface to enable students to ask their doubts easily. One preferred format for student queries is images, as it allows students to capture and post questions without typing complex equations and information. However, this format also presents difficulties, as images may contain multiple questions or textual noise that lowers the accuracy of existing single-query answering solutions. In this paper, we propose a method for extracting questions from text or images using a BERT-based deep learning model and compare it to the other rule-based and layout-based methods. Our method aims to improve the accuracy and efficiency of student query resolution in Edtech organizations.
Authors: Saurabh Saini, Kapil Ahuja, Siddartha Chennareddy, Karthik Boddupalli
Abstract: Cervical cancer stands as a predominant cause of female mortality, underscoring the need for regular screenings to enable early diagnosis and preemptive treatment of pre-cancerous conditions. The transformation zone in the cervix, where cellular differentiation occurs, plays a critical role in the detection of abnormalities. Colposcopy has emerged as a pivotal tool in cervical cancer prevention since it provides a meticulous examination of cervical abnormalities. However, challenges in visual evaluation necessitate the development of Computer Aided Diagnosis (CAD) systems. We propose a novel CAD system that combines the strengths of various deep-learning descriptors (ResNet50, ResNet101, and ResNet152) with appropriate feature normalization (min-max) as well as feature reduction technique (LDA). The combination of different descriptors ensures that all the features (low-level like edges and colour, high-level like shape and texture) are captured, feature normalization prevents biased learning, and feature reduction avoids overfitting. We do experiments on the IARC dataset provided by WHO. The dataset is initially segmented and balanced. Our approach achieves exceptional performance in the range of 97%-100% for both the normal-abnormal and the type classification. A competitive approach for type classification on the same dataset achieved 81%-91% performance.
Authors: Yassir Fathullah, Mark J. F. Gales
Abstract: Encoder-decoder foundation models have displayed state-of-the-art performance on a range of autoregressive sequence tasks. This paper proposes a simple and lightweight modification to such systems to control the behaviour according to a specific attribute of interest. This paper proposes a novel inference-efficient approach to modifying the behaviour of an encoder-decoder system according to a specific attribute of interest. Specifically, we show that a small proxy network can be used to find a sample-by-sample perturbation of the encoder output of a frozen foundation model to trigger the decoder to generate improved decodings. This work explores a specific realization of this framework focused on improving the COMET performance of Flan-T5 on Machine Translation and the WER of Whisper foundation models on Speech Recognition. Results display consistent improvements in performance evaluated through COMET and WER respectively. Furthermore, experiments also show that the proxies are robust to the exact nature of the data used to train them and can extend to other domains.
Authors: Ashish Anil Pawar, Vishnureddy Prashant Muskawar, Ritesh Tiku
Abstract: Algorithmic trading or Financial robots have been conquering the stock markets with their ability to fathom complex statistical trading strategies. But with the recent development of deep learning technologies, these strategies are becoming impotent. The DQN and A2C models have previously outperformed eminent humans in game-playing and robotics. In our work, we propose a reinforced portfolio manager offering assistance in the allocation of weights to assets. The environment proffers the manager the freedom to go long and even short on the assets. The weight allocation advisements are restricted to the choice of portfolio assets and tested empirically to knock benchmark indices. The manager performs financial transactions in a postulated liquid market without any transaction charges. This work provides the conclusion that the proposed portfolio manager with actions centered on weight allocations can surpass the risk-adjusted returns of conventional portfolio managers.
Authors: Jun Zhuang, Jack Cunningham, Chaowen Guan
Abstract: In the era of noisy intermediate-scale quantum (NISQ), variational quantum circuits (VQCs) have been widely applied in various domains, advancing the superiority of quantum circuits against classic models. Similar to classic models, regular VQCs can be optimized by various gradient-based methods. However, the optimization may be initially trapped in barren plateaus or eventually entangled in saddle points during training. These gradient issues can significantly undermine the trainability of VQC. In this work, we propose a strategy that regularizes model parameters with prior knowledge of the train data and Gaussian noise diffusion. We conduct ablation studies to verify the effectiveness of our strategy across four public datasets and demonstrate that our method can improve the trainability of VQCs against the above-mentioned gradient issues.
Authors: Chengqian Gao, William de Vazelhes, Hualin Zhang, Bin Gu, Zhiqiang Xu
Abstract: Evolution Strategies (ES) have emerged as a competitive alternative for model-free reinforcement learning, showcasing exemplary performance in tasks like Mujoco and Atari. Notably, they shine in scenarios with imperfect reward functions, making them invaluable for real-world applications where dense reward signals may be elusive. Yet, an inherent assumption in ES, that all input features are task-relevant, poses challenges, especially when confronted with irrelevant features common in real-world problems. This work scrutinizes this limitation, particularly focusing on the Natural Evolution Strategies (NES) variant. We propose NESHT, a novel approach that integrates Hard-Thresholding (HT) with NES to champion sparsity, ensuring only pertinent features are employed. Backed by rigorous analysis and empirical tests, NESHT demonstrates its promise in mitigating the pitfalls of irrelevant features and shines in complex decision-making problems like noisy Mujoco and Atari tasks.
Authors: Maksym Korablyov, Cheng-Hao Liu, Moksh Jain, Almer M. van der Sloot, Eric Jolicoeur, Edward Ruediger, Andrei Cristian Nica, Emmanuel Bengio, Kostiantyn Lapchevskyi, Daniel St-Cyr, Doris Alexandra Schuetz, Victor Ion Butoi, Jarrid Rector-Brooks, Simon Blackburn, Leo Feng, Hadi Nekoei, SaiKrishna Gottipati, Priyesh Vijayan, Prateek Gupta, Ladislav Ramp\'a\v{s}ek, Sasikanth Avancha, Pierre-Luc Bacon, William L. Hamilton, Brooks Paige, Sanchit Misra, Stanislaw Kamil Jastrzebski, Bharat Kaul, Doina Precup, Jos\'e Miguel Hern\'andez-Lobato, Marwin Segler, Michael Bronstein, Anne Marinier, Mike Tyers, Yoshua Bengio
Abstract: Despite substantial progress in machine learning for scientific discovery in recent years, truly de novo design of small molecules which exhibit a property of interest remains a significant challenge. We introduce LambdaZero, a generative active learning approach to search for synthesizable molecules. Powered by deep reinforcement learning, LambdaZero learns to search over the vast space of molecules to discover candidates with a desired property. We apply LambdaZero with molecular docking to design novel small molecules that inhibit the enzyme soluble Epoxide Hydrolase 2 (sEH), while enforcing constraints on synthesizability and drug-likeliness. LambdaZero provides an exponential speedup in terms of the number of calls to the expensive molecular docking oracle, and LambdaZero de novo designed molecules reach docking scores that would otherwise require the virtual screening of a hundred billion molecules. Importantly, LambdaZero discovers novel scaffolds of synthesizable, drug-like inhibitors for sEH. In in vitro experimental validation, a series of ligands from a generated quinazoline-based scaffold were synthesized, and the lead inhibitor N-(4,6-di(pyrrolidin-1-yl)quinazolin-2-yl)-N-methylbenzamide (UM0152893) displayed sub-micromolar enzyme inhibition of sEH.
Authors: Jayanth Shenoy, Xinjian Davis Zhang, Shlok Mehrotra, Bill Tao, Rem Yang, Han Zhao, Deepak Vasisht
Abstract: Satellite image time series (SITS) segmentation is crucial for many applications like environmental monitoring, land cover mapping and agricultural crop type classification. However, training models for SITS segmentation remains a challenging task due to the lack of abundant training data, which requires fine grained annotation. We propose S4 a new self-supervised pre-training approach that significantly reduces the requirement for labeled training data by utilizing two new insights: (a) Satellites capture images in different parts of the spectrum such as radio frequencies, and visible frequencies. (b) Satellite imagery is geo-registered allowing for fine-grained spatial alignment. We use these insights to formulate pre-training tasks in S4. We also curate m2s2-SITS, a large-scale dataset of unlabeled, spatially-aligned, multi-modal and geographic specific SITS that serves as representative pre-training data for S4. Finally, we evaluate S4 on multiple SITS segmentation datasets and demonstrate its efficacy against competing baselines while using limited labeled data.
Authors: Zhenjiang Mao, Dong-You Jhong, Ao Wang, Ivan Ruchkin
Abstract: Out-of-distribution (OOD) detection is essential in autonomous driving, to determine when learning-based components encounter unexpected inputs. Traditional detectors typically use encoder models with fixed settings, thus lacking effective human interaction capabilities. With the rise of large foundation models, multimodal inputs offer the possibility of taking human language as a latent representation, thus enabling language-defined OOD detection. In this paper, we use the cosine similarity of image and text representations encoded by the multimodal model CLIP as a new representation to improve the transparency and controllability of latent encodings used for visual anomaly detection. We compare our approach with existing pre-trained encoders that can only produce latent representations that are meaningless from the user's standpoint. Our experiments on realistic driving data show that the language-based latent representation performs better than the traditional representation of the vision encoder and helps improve the detection performance when combined with standard representations.
Authors: Guoping Xu, Xiaxia Wang, Xinglong Wu, Xuesong Leng, Yongchao Xu
Abstract: Deep learning has made significant progress in computer vision, specifically in image classification, object detection, and semantic segmentation. The skip connection has played an essential role in the architecture of deep neural networks,enabling easier optimization through residual learning during the training stage and improving accuracy during testing. Many neural networks have inherited the idea of residual learning with skip connections for various tasks, and it has been the standard choice for designing neural networks. This survey provides a comprehensive summary and outlook on the development of skip connections in deep neural networks. The short history of skip connections is outlined, and the development of residual learning in deep neural networks is surveyed. The effectiveness of skip connections in the training and testing stages is summarized, and future directions for using skip connections in residual learning are discussed. Finally, we summarize seminal papers, source code, models, and datasets that utilize skip connections in computer vision, including image classification, object detection, semantic segmentation, and image reconstruction. We hope this survey could inspire peer researchers in the community to develop further skip connections in various forms and tasks and the theory of residual learning in deep neural networks. The project page can be found at https://github.com/apple1986/Residual_Learning_For_Images
URLs: https://github.com/apple1986/Residual_Learning_For_Images
Authors: Guanyiman Fu, Fengchao Xiong, Jianfeng Lu, Jun Zhou, Yuntao Qian
Abstract: Denoising hyperspectral images (HSIs) is a crucial preprocessing procedure due to the noise originating from intra-imaging mechanisms and environmental factors. Utilizing domain-specific knowledge of HSIs, such as spectral correlation, spatial self-similarity, and spatial-spectral correlation, is essential for deep learning-based denoising. Existing methods are often constrained by running time, space complexity, and computational complexity, employing strategies that explore these priors separately. While the strategies can avoid some redundant information, considering that hyperspectral images are 3-D images with strong spatial continuity and spectral correlation, this kind of strategy inevitably overlooks subtle long-range spatial-spectral information that positively impacts image restoration. This paper proposes a Spatial-Spectral Selective State Space Model-based U-shaped network, termed Spatial-Spectral U-Mamba (SSUMamba), for hyperspectral image denoising. We can obtain complete global spatial-spectral correlation within a module thanks to the linear space complexity in State Space Model (SSM) computations. We introduce an Alternating Scan (SSAS) strategy for HSI data, which helps model the information flow in multiple directions in 3-D HSIs. Experimental results demonstrate that our method outperforms several compared methods. The source code will be available at https://github.com/lronkitty/SSUMamba.
Authors: Sanmitra Ghosh, Paul J. Birrell, Daniela De Angelis
Abstract: Likelihood-free inference methods based on neural conditional density estimation were shown to drastically reduce the simulation burden in comparison to classical methods such as ABC. When applied in the context of any latent variable model, such as a Hidden Markov model (HMM), these methods are designed to only estimate the parameters, rather than the joint distribution of the parameters and the hidden states. Naive application of these methods to a HMM, ignoring the inference of this joint posterior distribution, will thus produce an inaccurate estimate of the posterior predictive distribution, in turn hampering the assessment of goodness-of-fit. To rectify this problem, we propose a novel, sample-efficient likelihood-free method for estimating the high-dimensional hidden states of an implicit HMM. Our approach relies on learning directly the intractable posterior distribution of the hidden states, using an autoregressive-flow, by exploiting the Markov property. Upon evaluating our approach on some implicit HMMs, we found that the quality of the estimates retrieved using our method is comparable to what can be achieved using a much more computationally expensive SMC algorithm.
Authors: Xun Jiao, Fred Lin, Harish D. Dixit, Joel Coburn, Abhinav Pandey, Han Wang, Jianyu Huang, Venkat Ramesh, Wang Xu, Daniel Moore, Sriram Sankar
Abstract: Reliability of AI systems is a fundamental concern for the successful deployment and widespread adoption of AI technologies. Unfortunately, the escalating complexity and heterogeneity of AI hardware systems make them inevitably and increasingly susceptible to hardware faults (e.g., bit flips) that can potentially corrupt model parameters. Given this challenge, this paper aims to answer a critical question: How likely is a parameter corruption to result in an incorrect model output? To systematically answer this question, we propose a novel quantitative metric, Parameter Vulnerability Factor (PVF), inspired by architectural vulnerability factor (AVF) in computer architecture community, aiming to standardize the quantification of AI model resilience/vulnerability against parameter corruptions. We define a model parameter's PVF as the probability that a corruption in that particular model parameter will result in an incorrect output. Similar to AVF, this statistical concept can be derived from statistically extensive and meaningful fault injection (FI) experiments. In this paper, we present several use cases on applying PVF to three types of tasks/models during inference -- recommendation (DLRM), vision classification (CNN), and text classification (BERT). PVF can provide pivotal insights to AI hardware designers in balancing the tradeoff between fault protection and performance/efficiency such as mapping vulnerable AI parameter components to well-protected hardware modules. PVF metric is applicable to any AI model and has a potential to help unify and standardize AI vulnerability/resilience evaluation practice.
Authors: Shumaila Javaid, Nasir Saeed, Bin He
Abstract: Unmanned Aerial Vehicles (UAVs) have emerged as a transformative technology across diverse sectors, offering adaptable solutions to complex challenges in both military and civilian domains. Their expanding capabilities present a platform for further advancement by integrating cutting-edge computational tools like Artificial Intelligence (AI) and Machine Learning (ML) algorithms. These advancements have significantly impacted various facets of human life, fostering an era of unparalleled efficiency and convenience. Large Language Models (LLMs), a key component of AI, exhibit remarkable learning and adaptation capabilities within deployed environments, demonstrating an evolving form of intelligence with the potential to approach human-level proficiency. This work explores the significant potential of integrating UAVs and LLMs to propel the development of autonomous systems. We comprehensively review LLM architectures, evaluating their suitability for UAV integration. Additionally, we summarize the state-of-the-art LLM-based UAV architectures and identify novel opportunities for LLM embedding within UAV frameworks. Notably, we focus on leveraging LLMs to refine data analysis and decision-making processes, specifically for enhanced spectral sensing and sharing in UAV applications. Furthermore, we investigate how LLM integration expands the scope of existing UAV applications, enabling autonomous data processing, improved decision-making, and faster response times in emergency scenarios like disaster response and network restoration. Finally, we highlight crucial areas for future research that are critical for facilitating the effective integration of LLMs and UAVs.
Authors: Kota Kondo, Andrea Tagliabue, Xiaoyi Cai, Claudius Tewari, Olivia Garcia, Marcos Espitia-Alvarez, Jonathan P. How
Abstract: Traditional optimization-based planners, while effective, suffer from high computational costs, resulting in slow trajectory generation. A successful strategy to reduce computation time involves using Imitation Learning (IL) to develop fast neural network (NN) policies from those planners, which are treated as expert demonstrators. Although the resulting NN policies are effective at quickly generating trajectories similar to those from the expert, (1) their output does not explicitly account for dynamic feasibility, and (2) the policies do not accommodate changes in the constraints different from those used during training. To overcome these limitations, we propose Constraint-Guided Diffusion (CGD), a novel IL-based approach to trajectory planning. CGD leverages a hybrid learning/online optimization scheme that combines diffusion policies with a surrogate efficient optimization problem, enabling the generation of collision-free, dynamically feasible trajectories. The key ideas of CGD include dividing the original challenging optimization problem solved by the expert into two more manageable sub-problems: (a) efficiently finding collision-free paths, and (b) determining a dynamically-feasible time-parametrization for those paths to obtain a trajectory. Compared to conventional neural network architectures, we demonstrate through numerical evaluations significant improvements in performance and dynamic feasibility under scenarios with new constraints never encountered during training.
Authors: Han Wang, Eiji Kawasaki, Guillaume Damblin, Geoffrey Daniel
Abstract: We present new Bayesian Last Layer models in the setting of multivariate regression under heteroscedastic noise, and propose an optimization algorithm for parameter learning. Bayesian Last Layer combines Bayesian modelling of the predictive distribution with neural networks for parameterization of the prior, and has the attractive property of uncertainty quantification with a single forward pass. The proposed framework is capable of disentangling the aleatoric and epistemic uncertainty, and can be used to transfer a canonically trained deep neural network to new data domains with uncertainty-aware capability.
Authors: Jian Meng, Yuan Liao, Anupreetham Anupreetham, Ahmed Hasssan, Shixing Yu, Han-sok Suh, Xiaofeng Hu, Jae-sun Seo
Abstract: The development of model compression is continuously motivated by the evolution of various neural network accelerators with ASIC or FPGA. On the algorithm side, the ultimate goal of quantization or pruning is accelerating the expensive DNN computations on low-power hardware. However, such a "design-and-deploy" workflow faces under-explored challenges in the current hardware-algorithm co-design community. First, although the state-of-the-art quantization algorithm can achieve low precision with negligible degradation of accuracy, the latest deep learning framework (e.g., PyTorch) can only support non-customizable 8-bit precision, data format, and parameter extraction. Secondly, the objective of quantization is to enable the computation with low-precision data. However, the current SoTA algorithm treats the quantized integer as an intermediate result, while the final output of the quantizer is the "discretized" floating-point values, ignoring the practical needs and adding additional workload to hardware designers for integer parameter extraction and layer fusion. Finally, the compression toolkits designed by the industry are constrained to their in-house product or a handful of algorithms. The limited degree of freedom in the current toolkit and the under-explored customization hinder the prototype ASIC or FPGA-based accelerator design. To resolve these challenges, we propose Torch2Chip, an open-sourced, fully customizable, and high-performance toolkit that supports user-designed compression followed by automatic model fusion and parameter extraction. Torch2Chip incorporates the hierarchical design workflow, and the user-customized compression algorithm will be directly packed into the deployment-ready format for prototype chip verification with either CNN or vision transformer (ViT). The code is available at https://github.com/SeoLabCornell/torch2chip.
Authors: Leon Eisemann, Mirjam Fehling-Kaschek, Henrik Gommel, David Hermann, Marvin Klemp, Martin Lauer, Benjamin Lickert, Florian Luettner, Robin Moss, Nicole Neis, Maria Pohle, Simon Romanski, Daniel Stadler, Alexander Stolz, Jens Ziehn, Jingxing Zhou
Abstract: With growing complexity and criticality of automated driving functions in road traffic and their operational design domains (ODD), there is increasing demand for covering significant proportions of development, validation, and verification in virtual environments and through simulation models. If, however, simulations are meant not only to augment real-world experiments, but to replace them, quantitative approaches are required that measure to what degree and under which preconditions simulation models adequately represent reality, and thus, using their results accordingly. Especially in R&D areas related to the safety impact of the "open world", there is a significant shortage of real-world data to parameterize and/or validate simulations - especially with respect to the behavior of human traffic participants, whom automated driving functions will meet in mixed traffic. We present an approach to systematically acquire data in public traffic by heterogeneous means, transform it into a unified representation, and use it to automatically parameterize traffic behavior models for use in data-driven virtual validation of automated driving functions.
Authors: Joonho Lee, Marko Bjelonic, Alexander Reske, Lorenz Wellhausen, Takahiro Miki, Marco Hutter
Abstract: Autonomous wheeled-legged robots have the potential to transform logistics systems, improving operational efficiency and adaptability in urban environments. Navigating urban environments, however, poses unique challenges for robots, necessitating innovative solutions for locomotion and navigation. These challenges include the need for adaptive locomotion across varied terrains and the ability to navigate efficiently around complex dynamic obstacles. This work introduces a fully integrated system comprising adaptive locomotion control, mobility-aware local navigation planning, and large-scale path planning within the city. Using model-free reinforcement learning (RL) techniques and privileged learning, we develop a versatile locomotion controller. This controller achieves efficient and robust locomotion over various rough terrains, facilitated by smooth transitions between walking and driving modes. It is tightly integrated with a learned navigation controller through a hierarchical RL framework, enabling effective navigation through challenging terrain and various obstacles at high speed. Our controllers are integrated into a large-scale urban navigation system and validated by autonomous, kilometer-scale navigation missions conducted in Zurich, Switzerland, and Seville, Spain. These missions demonstrate the system's robustness and adaptability, underscoring the importance of integrated control systems in achieving seamless navigation in complex environments. Our findings support the feasibility of wheeled-legged robots and hierarchical RL for autonomous navigation, with implications for last-mile delivery and beyond.
Authors: Tian Xie, Xueru Zhang
Abstract: This paper studies algorithmic decision-making in the presence of strategic individual behaviors, where an ML model is used to make decisions about human agents and the latter can adapt their behavior strategically to improve their future data. Existing results on strategic learning have largely focused on the linear setting where agents with linear labeling functions best respond to a (noisy) linear decision policy. Instead, this work focuses on general non-linear settings where agents respond to the decision policy with only "local information" of the policy. Moreover, we simultaneously consider the objectives of maximizing decision-maker welfare (model prediction accuracy), social welfare (agent improvement caused by strategic behaviors), and agent welfare (the extent that ML underestimates the agents). We first generalize the agent best response model in previous works to the non-linear setting, then reveal the compatibility of welfare objectives. We show the three welfare can attain the optimum simultaneously only under restrictive conditions which are challenging to achieve in non-linear settings. The theoretical results imply that existing works solely maximizing the welfare of a subset of parties inevitably diminish the welfare of the others. We thus claim the necessity of balancing the welfare of each party in non-linear settings and propose an irreducible optimization algorithm suitable for general strategic learning. Experiments on synthetic and real data validate the proposed algorithm.
Authors: Tanya Chowdhury, Yair Zick, James Allan
Abstract: Several works propose various post-hoc, model-agnostic explanations for the task of ranking, i.e. the task of ordering a set of documents, via feature attribution methods. However, these attributions are seen to weakly correlate and sometimes contradict each other. In classification/regression, several works focus on \emph{axiomatic characterization} of feature attribution methods, showing that a certain method uniquely satisfies a set of desirable properties. However, no such efforts have been taken in the space of feature attributions for the task of ranking. We take an axiomatic game-theoretic approach, popular in the feature attribution community, to identify candidate attribution methods for ranking tasks. We first define desirable axioms: Rank-Efficiency, Rank-Missingness, Rank-Symmetry and Rank-Monotonicity, all variants of the classical Shapley axioms. Next, we introduce Rank-SHAP, a feature attribution algorithm for the general ranking task, which is an extension to classical Shapley values. We identify a polynomial-time algorithm for computing approximate Rank-SHAP values and evaluate the computational efficiency and accuracy of our algorithm under various scenarios. We also evaluate its alignment with human intuition with a user study. Lastly, we theoretically examine popular rank attribution algorithms, EXS and Rank-LIME, and evaluate their capacity to satisfy the classical Shapley axioms.
Authors: Sairamvinay Vijayaraghavan, Prasant Mohapatra
Abstract: Explainable Recommendation has been gaining attention over the last few years in industry and academia. Explanations provided along with recommendations in a recommender system framework have many uses: particularly reasoning why a suggestion is provided and how well an item aligns with a user's personalized preferences. Hence, explanations can play a huge role in influencing users to purchase products. However, the reliability of the explanations under varying scenarios has not been strictly verified from an empirical perspective. Unreliable explanations can bear strong consequences such as attackers leveraging explanations for manipulating and tempting users to purchase target items that the attackers would want to promote. In this paper, we study the vulnerability of existent feature-oriented explainable recommenders, particularly analyzing their performance under different levels of external noises added into model parameters. We conducted experiments by analyzing three important state-of-the-art (SOTA) explainable recommenders when trained on two widely used e-commerce based recommendation datasets of different scales. We observe that all the explainable models are vulnerable to increased noise levels. Experimental results verify our hypothesis that the ability to explain recommendations does decrease along with increasing noise levels and particularly adversarial noise does contribute to a much stronger decrease. Our study presents an empirical verification on the topic of robust explanations in recommender systems which can be extended to different types of explainable recommenders in RS.
Authors: Sairamvinay Vijayaraghavan, Prasant Mohapatra
Abstract: Explainable Recommender Systems is an important field of study which provides reasons behind the suggested recommendations. Explanations with recommender systems are useful for developers while debugging anomalies within the system and for consumers while interpreting the model's effectiveness in capturing their true preferences towards items. However, most of the existing state-of-the-art (SOTA) explainable recommenders could not retain their explanation capability under noisy circumstances and moreover are not generalizable across different datasets. The robustness of the explanations must be ensured so that certain malicious attackers do not manipulate any high-stake decision scenarios to their advantage, which could cause severe consequences affecting large groups of interest. In this work, we present a general framework for feature-aware explainable recommenders that can withstand external attacks and provide robust and generalized explanations. This paper presents a novel framework which could be utilized as an additional defense tool, preserving the global explainability when subject to model-based white box attacks. Our framework is simple to implement and supports different methods regardless of the internal model structure and intrinsic utility within any model. We experimented our framework on two architecturally different feature-based SOTA explainable algorithms by training them on three popular e-commerce datasets of increasing scales. We noticed that both the algorithms displayed an overall improvement in the quality and robustness of the global explainability under normal as well as noisy environments across all the datasets, indicating the flexibility and mutability of our framework.
Authors: Riley Simmons-Edler, Ryan Badman, Shayne Longpre, Kanaka Rajan
Abstract: The recent embrace of machine learning (ML) in the development of autonomous weapons systems (AWS) creates serious risks to geopolitical stability and the free exchange of ideas in AI research. This topic has received comparatively little attention of late compared to risks stemming from superintelligent artificial general intelligence (AGI), but requires fewer assumptions about the course of technological development and is thus a nearer-future issue. ML is already enabling the substitution of AWS for human soldiers in many battlefield roles, reducing the upfront human cost, and thus political cost, of waging offensive war. In the case of peer adversaries, this increases the likelihood of "low intensity" conflicts which risk escalation to broader warfare. In the case of non-peer adversaries, it reduces the domestic blowback to wars of aggression. This effect can occur regardless of other ethical issues around the use of military AI such as the risk of civilian casualties, and does not require any superhuman AI capabilities. Further, the military value of AWS raises the specter of an AI-powered arms race and the misguided imposition of national security restrictions on AI research. Our goal in this paper is to raise awareness among the public and ML researchers on the near-future risks posed by full or near-full autonomy in military technology, and we provide regulatory suggestions to mitigate these risks. We call upon AI policy experts and the defense AI community in particular to embrace transparency and caution in their development and deployment of AWS to avoid the negative effects on global stability and AI research that we highlight here.
Authors: Md Robiul Islam, Al Amin, Aniqua Nusrat Zereen
Abstract: Texting stands out as the most prominent form of communication worldwide. Individual spend significant amount of time writing whole texts to send emails or write something on social media, which is time consuming in this modern era. Word prediction and sentence completion will be suitable and appropriate in the Bangla language to make textual information easier and more convenient. This paper expands the scope of Bangla language processing by introducing a Bi-LSTM model that effectively handles Bangla next-word prediction and Bangla sentence generation, demonstrating its versatility and potential impact. We proposed a new Bi-LSTM model to predict a following word and complete a sentence. We constructed a corpus dataset from various news portals, including bdnews24, BBC News Bangla, and Prothom Alo. The proposed approach achieved superior results in word prediction, reaching 99\% accuracy for both 4-gram and 5-gram word predictions. Moreover, it demonstrated significant improvement over existing methods, achieving 35\%, 75\%, and 95\% accuracy for uni-gram, bi-gram, and tri-gram word prediction, respectively
Authors: Xue Wen Tan, Stanley Kok
Abstract: Every publicly traded company in the US is required to file an annual 10-K financial report, which contains a wealth of information about the company. In this paper, we propose an explainable deep-learning model, called FinBERT-XRC, that takes a 10-K report as input, and automatically assesses the post-event return volatility risk of its associated company. In contrast to previous systems, our proposed model simultaneously offers explanations of its classification decision at three different levels: the word, sentence, and corpus levels. By doing so, our model provides a comprehensive interpretation of its prediction to end users. This is particularly important in financial domains, where the transparency and accountability of algorithmic predictions play a vital role in their application to decision-making processes. Aside from its novel interpretability, our model surpasses the state of the art in predictive accuracy in experiments on a large real-world dataset of 10-K reports spanning six years.
Authors: Miruna Be\c{t}ianu, Abele M\u{a}lan, Marco Aldinucci, Robert Birke, Lydia Chen
Abstract: Large language models (LLMs) increasingly serve as the backbone for classifying text associated with distinct domains and simultaneously several labels (classes). When encountering domain shifts, e.g., classifier of movie reviews from IMDb to Rotten Tomatoes, adapting such an LLM-based multi-label classifier is challenging due to incomplete label sets at the target domain and daunting training overhead. The existing domain adaptation methods address either image multi-label classifiers or text binary classifiers. In this paper, we design DALLMi, Domain Adaptation Large Language Model interpolator, a first-of-its-kind semi-supervised domain adaptation method for text data models based on LLMs, specifically BERT. The core of DALLMi is the novel variation loss and MixUp regularization, which jointly leverage the limited positively labeled and large quantity of unlabeled text and, importantly, their interpolation from the BERT word embeddings. DALLMi also introduces a label-balanced sampling strategy to overcome the imbalance between labeled and unlabeled data. We evaluate DALLMi against the partial-supervised and unsupervised approach on three datasets under different scenarios of label availability for the target domain. Our results show that DALLMi achieves higher mAP than unsupervised and partially-supervised approaches by 19.9% and 52.2%, respectively.
Authors: Changliang Zhou, Xi Lin, Zhenkun Wang, Xialiang Tong, Mingxuan Yuan, Qingfu Zhang
Abstract: The neural combinatorial optimization (NCO) approach has shown great potential for solving routing problems without the requirement of expert knowledge. However, existing constructive NCO methods cannot directly solve large-scale instances, which significantly limits their application prospects. To address these crucial shortcomings, this work proposes a novel Instance-Conditioned Adaptation Model (ICAM) for better large-scale generalization of neural combinatorial optimization. In particular, we design a powerful yet lightweight instance-conditioned adaptation module for the NCO model to generate better solutions for instances across different scales. In addition, we develop an efficient three-stage reinforcement learning-based training scheme that enables the model to learn cross-scale features without any labeled optimal solution. Experimental results show that our proposed method is capable of obtaining excellent results with a very fast inference time in solving Traveling Salesman Problems (TSPs) and Capacitated Vehicle Routing Problems (CVRPs) across different scales. To the best of our knowledge, our model achieves state-of-the-art performance among all RL-based constructive methods for TSP and CVRP with up to 1,000 nodes.
Authors: Firuz Juraev, Mohammed Abuhamad, Simon S. Woo, George K Thiruvathukal, Tamer Abuhmed
Abstract: Rapid advancements of deep learning are accelerating adoption in a wide variety of applications, including safety-critical applications such as self-driving vehicles, drones, robots, and surveillance systems. These advancements include applying variations of sophisticated techniques that improve the performance of models. However, such models are not immune to adversarial manipulations, which can cause the system to misbehave and remain unnoticed by experts. The frequency of modifications to existing deep learning models necessitates thorough analysis to determine the impact on models' robustness. In this work, we present an experimental evaluation of the effects of model modifications on deep learning model robustness using adversarial attacks. Our methodology involves examining the robustness of variations of models against various adversarial attacks. By conducting our experiments, we aim to shed light on the critical issue of maintaining the reliability and safety of deep learning models in safety- and security-critical applications. Our results indicate the pressing demand for an in-depth assessment of the effects of model changes on the robustness of models.
Authors: Zhiyu Guo, Hidetaka Kamigaito, Taro Wanatnabe
Abstract: The rapid advancement in Large Language Models (LLMs) has markedly enhanced the capabilities of language understanding and generation. However, the substantial model size poses hardware challenges, affecting both memory size for serving and inference latency for token generation. To address those challenges, we propose Dependency-aware Semi-structured Sparsity (DaSS), a novel method for the recent prevalent SwiGLU-based LLMs pruning. Our approach incorporates structural dependency into the weight magnitude-based unstructured pruning. We introduce an MLP-specific pruning metric that evaluates the importance of each weight by jointly considering its magnitude and its corresponding MLP intermediate activation norms. DaSS facilitates a balance between the adaptability offered by unstructured pruning and the structural consistency inherent in dependency-based structured pruning. Empirical evaluations on Mistral and LLaMA2 model families demonstrate that DaSS not only outperforms both SparseGPT and Wanda in achieving hardware-friendly N:M sparsity patterns but also maintains the computational efficiency of Wanda.
Authors: Weigutian Ou, Philipp Schenkel, Helmut B\"olcskei
Abstract: We establish the fundamental limits in the approximation of Lipschitz functions by deep ReLU neural networks with finite-precision weights. Specifically, three regimes, namely under-, over-, and proper quantization, in terms of minimax approximation error behavior as a function of network weight precision, are identified. This is accomplished by deriving nonasymptotic tight lower and upper bounds on the minimax approximation error. Notably, in the proper-quantization regime, neural networks exhibit memory-optimality in the approximation of Lipschitz functions. Deep networks have an inherent advantage over shallow networks in achieving memory-optimality. We also develop the notion of depth-precision tradeoff, showing that networks with high-precision weights can be converted into functionally equivalent deeper networks with low-precision weights, while preserving memory-optimality. This idea is reminiscent of sigma-delta analog-to-digital conversion, where oversampling rate is traded for resolution in the quantization of signal samples. We improve upon the best-known ReLU network approximation results for Lipschitz functions and describe a refinement of the bit extraction technique which could be of independent general interest.
Authors: Firuz Juraev, Mohammed Abuhamad, Eric Chan-Tin, George K. Thiruvathukal, Tamer Abuhmed
Abstract: Deep Learning (DL) is rapidly maturing to the point that it can be used in safety- and security-crucial applications. However, adversarial samples, which are undetectable to the human eye, pose a serious threat that can cause the model to misbehave and compromise the performance of such applications. Addressing the robustness of DL models has become crucial to understanding and defending against adversarial attacks. In this study, we perform comprehensive experiments to examine the effect of adversarial attacks and defenses on various model architectures across well-known datasets. Our research focuses on black-box attacks such as SimBA, HopSkipJump, MGAAttack, and boundary attacks, as well as preprocessor-based defensive mechanisms, including bits squeezing, median smoothing, and JPEG filter. Experimenting with various models, our results demonstrate that the level of noise needed for the attack increases as the number of layers increases. Moreover, the attack success rate decreases as the number of layers increases. This indicates that model complexity and robustness have a significant relationship. Investigating the diversity and robustness relationship, our experiments with diverse models show that having a large number of parameters does not imply higher robustness. Our experiments extend to show the effects of the training dataset on model robustness. Using various datasets such as ImageNet-1000, CIFAR-100, and CIFAR-10 are used to evaluate the black-box attacks. Considering the multiple dimensions of our analysis, e.g., model complexity and training dataset, we examined the behavior of black-box attacks when models apply defenses. Our results show that applying defense strategies can significantly reduce attack effectiveness. This research provides in-depth analysis and insight into the robustness of DL models against various attacks, and defenses.
Authors: Patrik Reizinger, Szilvia Ujv\'ary, Anna M\'esz\'aros, Anna Kerekes, Wieland Brendel, Ferenc Husz\'ar
Abstract: The last decade has seen blossoming research in deep learning theory attempting to answer, "Why does deep learning generalize?" A powerful shift in perspective precipitated this progress: the study of overparametrized models in the interpolation regime. In this paper, we argue that another perspective shift is due, since some of the desirable qualities of LLMs are not a consequence of good statistical generalization and require a separate theoretical explanation. Our core argument relies on the observation that AR probabilistic models are inherently non-identifiable: models zero or near-zero KL divergence apart -- thus, equivalent test loss -- can exhibit markedly different behaviors. We support our position with mathematical examples and empirical observations, illustrating why non-identifiability has practical relevance through three case studies: (1) the non-identifiability of zero-shot rule extrapolation; (2) the approximate non-identifiability of in-context learning; and (3) the non-identifiability of fine-tunability. We review promising research directions focusing on LLM-relevant generalization measures, transferability, and inductive biases.
Authors: Rasoul Najafi Koopas, Shahed Rezaei, Natalie Rauter, Richard Ostwald, Rolf Lammering
Abstract: In this study, we develop a novel multi-fidelity deep learning approach that transforms low-fidelity solution maps into high-fidelity ones by incorporating parametric space information into a standard autoencoder architecture. It is shown that, due to the integration of parametric space data, this method requires significantly less training data to achieve effective performance in predicting high-fidelity solution from the low-fidelity one. In this study, our focus is on a 2D steady-state heat transfer analysis in highly heterogeneous materials microstructure, where the spatial distribution of heat conductivity coefficients for two distinct materials is condensed. Subsequently, the boundary value problem is solved on the coarsest grid using a pre-trained physics-informed neural operator network. Afterward, the calculated low-fidelity result is upscaled using the newly designed enhanced autoencoder. The novelty of the developed enhanced autoencoder lies in the concatenation of heat conductivity maps of different resolutions to the decoder segment in distinct steps. We then compare the outcomes of developed algorithm with the corresponding finite element results, standard U-Net architecture as well as other upscaling approaches such as interpolation functions of varying orders and feedforward neural networks (FFNN). The analysis of the results based on the new approach demonstrates superior performance compared to other approaches in terms of computational cost and error on the test cases. Therefore, as a potential supplement to neural operators networks, our architecture upscales low-fidelity solutions to high-fidelity ones while preserving critical details that are often lost in conventional upscaling methods, especially at sharp interfaces, such as those encountered with interpolation methods.
Authors: Margarida M. Campos, Ant\'onio Farinhas, Chrysoula Zerva, M\'ario A. T. Figueiredo, Andr\'e F. T. Martins
Abstract: The rapid proliferation of large language models and natural language processing (NLP) applications creates a crucial need for uncertainty quantification to mitigate risks such as hallucinations and to enhance decision-making reliability in critical applications. Conformal prediction is emerging as a theoretically sound and practically useful framework, combining flexibility with strong statistical guarantees. Its model-agnostic and distribution-free nature makes it particularly promising to address the current shortcomings of NLP systems that stem from the absence of uncertainty quantification. This paper provides a comprehensive survey of conformal prediction techniques, their guarantees, and existing applications in NLP, pointing to directions for future research and open challenges.
Authors: Lea Schaab, Anna Kruspe
Abstract: Sentiment or mood can express themselves on various levels in music. In automatic analysis, the actual audio data is usually analyzed, but the lyrics can also play a crucial role in the perception of moods. We first evaluate various models for sentiment analysis based on lyrics and audio separately. The corresponding approaches already show satisfactory results, but they also exhibit weaknesses, the causes of which we examine in more detail. Furthermore, different approaches to combining the audio and lyrics results are proposed and evaluated. Considering both modalities generally leads to improved performance. We investigate misclassifications and (also intentional) contradictions between audio and lyrics sentiment more closely, and identify possible causes. Finally, we address fundamental problems in this research area, such as high subjectivity, lack of data, and inconsistency in emotion taxonomies.
Authors: Patrick Saux
Abstract: This thesis aims to study some of the mathematical challenges that arise in the analysis of statistical sequential decision-making algorithms for postoperative patients follow-up. Stochastic bandits (multiarmed, contextual) model the learning of a sequence of actions (policy) by an agent in an uncertain environment in order to maximise observed rewards. To learn optimal policies, bandit algorithms have to balance the exploitation of current knowledge and the exploration of uncertain actions. Such algorithms have largely been studied and deployed in industrial applications with large datasets, low-risk decisions and clear modelling assumptions, such as clickthrough rate maximisation in online advertising. By contrast, digital health recommendations call for a whole new paradigm of small samples, risk-averse agents and complex, nonparametric modelling. To this end, we developed new safe, anytime-valid concentration bounds, (Bregman, empirical Chernoff), introduced a new framework for risk-aware contextual bandits (with elicitable risk measures) and analysed a novel class of nonparametric bandit algorithms under weak assumptions (Dirichlet sampling). In addition to the theoretical guarantees, these results are supported by in-depth empirical evidence. Finally, as a first step towards personalised postoperative follow-up recommendations, we developed with medical doctors and surgeons an interpretable machine learning model to predict the long-term weight trajectories of patients after bariatric surgery.
Authors: Nicolas Dewolf
Abstract: In the past decades, most work in the area of data analysis and machine learning was focused on optimizing predictive models and getting better results than what was possible with existing models. To what extent the metrics with which such improvements were measured were accurately capturing the intended goal, whether the numerical differences in the resulting values were significant, or whether uncertainty played a role in this study and if it should have been taken into account, was of secondary importance. Whereas probability theory, be it frequentist or Bayesian, used to be the gold standard in science before the advent of the supercomputer, it was quickly replaced in favor of black box models and sheer computing power because of their ability to handle large data sets. This evolution sadly happened at the expense of interpretability and trustworthiness. However, while people are still trying to improve the predictive power of their models, the community is starting to realize that for many applications it is not so much the exact prediction that is of importance, but rather the variability or uncertainty. The work in this dissertation tries to further the quest for a world where everyone is aware of uncertainty, of how important it is and how to embrace it instead of fearing it. A specific, though general, framework that allows anyone to obtain accurate uncertainty estimates is singled out and analysed. Certain aspects and applications of the framework -- dubbed `conformal prediction' -- are studied in detail. Whereas many approaches to uncertainty quantification make strong assumptions about the data, conformal prediction is, at the time of writing, the only framework that deserves the title `distribution-free'. No parametric assumptions have to be made and the nonparametric results also hold without having to resort to the law of large numbers in the asymptotic regime.
Authors: Niclas F\"uhrling, Kengo Ando, Giuseppe Thadeu Freitas de Abreu, David Gonz\'alez G., Osvaldo Gonsa
Abstract: We consider a novel algorithm, for the completion of partially observed low-rank matrices in a structured setting where each entry can be chosen from a finite discrete alphabet set, such as in common recommender systems. The proposed low-rank matrix completion (MC) method is an improved variation of state-of-the-art (SotA) discrete aware matrix completion method which we previously proposed, in which discreteness is enforced by an $\ell_0$-norm regularizer, not by replaced with the $\ell_1$-norm, but instead approximated by a continuous and differentiable function normalized via fractional programming (FP) under a proximal gradient (PG) framework. Simulation results demonstrate the superior performance of the new method compared to the SotA techniques as well as the earlier $\ell_1$-norm-based discrete-aware matrix completion approach.
Authors: Denise Moussa, Germans Hirsch, Christian Riess
Abstract: Audio recordings may provide important evidence in criminal investigations. One such case is the forensic association of the recorded audio to the recording location. For example, a voice message may be the only investigative cue to narrow down the candidate sites for a crime. Up to now, several works provide tools for closed-set recording environment classification under relatively clean recording conditions. However, in forensic investigations, the candidate locations are case-specific. Thus, closed-set tools are not applicable without retraining on a sufficient amount of training samples for each case and respective candidate set. In addition, a forensic tool has to deal with audio material from uncontrolled sources with variable properties and quality. In this work, we therefore attempt a major step towards practical forensic application scenarios. We propose a representation learning framework called EnvId, short for environment identification. EnvId avoids case-specific retraining. Instead, it is the first tool for robust few-shot classification of unseen environment locations. We demonstrate that EnvId can handle forensically challenging material. It provides good quality predictions even under unseen signal degradations, environment characteristics or recording position mismatches. Our code and datasets will be made publicly available upon acceptance.
Authors: No\'e Tits, Prernna Bhatnagar, Thierry Dutoit
Abstract: In this paper, we present a novel approach for text independent phone-to-audio alignment based on phoneme recognition, representation learning and knowledge transfer. Our method leverages a self-supervised model (wav2vec2) fine-tuned for phoneme recognition using a Connectionist Temporal Classification (CTC) loss, a dimension reduction model and a frame-level phoneme classifier trained thanks to forced-alignment labels (using Montreal Forced Aligner) to produce multi-lingual phonetic representations, thus requiring minimal additional training. We evaluate our model using synthetic native data from the TIMIT dataset and the SCRIBE dataset for American and British English, respectively. Our proposed model outperforms the state-of-the-art (charsiu) in statistical metrics and has applications in language learning and speech processing systems. We leave experiments on other languages for future work but the design of the system makes it easily adaptable to other languages.
Authors: Olivier Jeunen, Jatin Mandav, Ivan Potapov, Nakul Agarwal, Sourabh Vaid, Wenzhe Shi, Aleksei Ustimenko
Abstract: Real-world recommender systems often need to balance multiple objectives when deciding which recommendations to present to users. These include behavioural signals (e.g. clicks, shares, dwell time), as well as broader objectives (e.g. diversity, fairness). Scalarisation methods are commonly used to handle this balancing task, where a weighted average of per-objective reward signals determines the final score used for ranking. Naturally, how these weights are computed exactly, is key to success for any online platform. We frame this as a decision-making task, where the scalarisation weights are actions taken to maximise an overall North Star reward (e.g. long-term user retention or growth). We extend existing policy learning methods to the continuous multivariate action domain, proposing to maximise a pessimistic lower bound on the North Star reward that the learnt policy will yield. Typical lower bounds based on normal approximations suffer from insufficient coverage, and we propose an efficient and effective policy-dependent correction for this. We provide guidance to design stochastic data collection policies, as well as highly sensitive reward signals. Empirical observations from simulations, offline and online experiments highlight the efficacy of our deployed approach.
Authors: Lu\'is Esp\'irito Santo, Geraint Wiggins, Am\'ilcar Cardoso
Abstract: Formalizing creativity-related concepts has been a long-term goal of Computational Creativity. To the same end, we explore Formal Learning Theory in the context of creativity. We provide an introduction to the main concepts of this framework and a re-interpretation of terms commonly found in creativity discussions, proposing formal definitions for novelty and transformational creativity. This formalisation marks the beginning of a research branch we call Formal Creativity Theory, exploring how learning can be included as preparation for exploratory behaviour and how learning is a key part of transformational creative behaviour. By employing these definitions, we argue that, while novelty is neither necessary nor sufficient for transformational creativity in general, when using an inspiring set, rather than a sequence of experiences, an agent actually requires novelty for transformational creativity to occur.
Authors: Hsuvas Borkakoty, Luis Espinosa-Anke
Abstract: Hoaxes are a recognised form of disinformation created deliberately, with potential serious implications in the credibility of reference knowledge resources such as Wikipedia. What makes detecting Wikipedia hoaxes hard is that they often are written according to the official style guidelines. In this work, we first provide a systematic analysis of the similarities and discrepancies between legitimate and hoax Wikipedia articles, and introduce Hoaxpedia, a collection of 311 Hoax articles (from existing literature as well as official Wikipedia lists) alongside semantically similar real articles. We report results of binary classification experiments in the task of predicting whether a Wikipedia article is real or hoax, and analyze several settings as well as a range of language models. Our results suggest that detecting deceitful content in Wikipedia based on content alone, despite not having been explored much in the past, is a promising direction.
Authors: Sang Bin Moon, Abolfazl Hashemi
Abstract: The Adversarial Markov Decision Process (AMDP) is a learning framework that deals with unknown and varying tasks in decision-making applications like robotics and recommendation systems. A major limitation of the AMDP formalism, however, is pessimistic regret analysis results in the sense that although the cost function can change from one episode to the next, the evolution in many settings is not adversarial. To address this, we introduce and study a new variant of AMDP, which aims to minimize regret while utilizing a set of cost predictors. For this setting, we develop a new policy search method that achieves a sublinear optimistic regret with high probability, that is a regret bound which gracefully degrades with the estimation power of the cost predictors. Establishing such optimistic regret bounds is nontrivial given that (i) as we demonstrate, the existing importance-weighted cost estimators cannot establish optimistic bounds, and (ii) the feedback model of AMDP is different (and more realistic) than the existing optimistic online learning works. Our result, in particular, hinges upon developing a novel optimistically biased cost estimator that leverages cost predictors and enables a high-probability regret analysis without imposing restrictive assumptions. We further discuss practical extensions of the proposed scheme and demonstrate its efficacy numerically.
Authors: Yijun Yan, Jinchang Ren, Barry Harrison, Oliver Lewis, Yinhe Li, Ping Ma
Abstract: Peat, a crucial component in whisky production, imparts distinctive and irreplaceable flavours to the final product. However, the extraction of peat disrupts ancient ecosystems and releases significant amounts of carbon, contributing to climate change. This paper aims to address this issue by conducting a feasibility study on enhancing peat use efficiency in whisky manufacturing through non-destructive analysis using hyperspectral imaging. Results show that shot-wave infrared (SWIR) data is more effective for analyzing peat samples and predicting total phenol levels, with accuracies up to 99.81%.
Authors: Ghalyah H. Aleryani, Wael Deabes, Khaled Albishre, Alaa E. Abdel-Hakim
Abstract: The complex challenge of detecting sarcasm in Arabic speech on social media is increased by the language diversity and the nature of sarcastic expressions. There is a significant gap in the capability of existing models to effectively interpret sarcasm in Arabic, which mandates the necessity for more sophisticated and precise detection methods. In this paper, we investigate the impact of a fundamental preprocessing component on sarcasm speech detection. While emojis play a crucial role in mitigating the absence effect of body language and facial expressions in modern communication, their impact on automated text analysis, particularly in sarcasm detection, remains underexplored. We investigate the impact of emoji exclusion from datasets on the performance of sarcasm detection models in social media content for Arabic as a vocabulary-super rich language. This investigation includes the adaptation and enhancement of AraBERT pre-training models, specifically by excluding emojis, to improve sarcasm detection capabilities. We use AraBERT pre-training to refine the specified models, demonstrating that the removal of emojis can significantly boost the accuracy of sarcasm detection. This approach facilitates a more refined interpretation of language, eliminating the potential confusion introduced by non-textual elements. The evaluated AraBERT models, through the focused strategy of emoji removal, adeptly navigate the complexities of Arabic sarcasm. This study establishes new benchmarks in Arabic natural language processing and presents valuable insights for social media platforms.
Authors: Peter Schmitt-F\"orster, Tobias Sutter
Abstract: We propose a new Q-learning variant, called 2RA Q-learning, that addresses some weaknesses of existing Q-learning methods in a principled manner. One such weakness is an underlying estimation bias which cannot be controlled and often results in poor performance. We propose a distributionally robust estimator for the maximum expected value term, which allows us to precisely control the level of estimation bias introduced. The distributionally robust estimator admits a closed-form solution such that the proposed algorithm has a computational cost per iteration comparable to Watkins' Q-learning. For the tabular case, we show that 2RA Q-learning converges to the optimal policy and analyze its asymptotic mean-squared error. Lastly, we conduct numerical experiments for various settings, which corroborate our theoretical findings and indicate that 2RA Q-learning often performs better than existing methods.
Authors: Michael R. Lyu, Baishakhi Ray, Abhik Roychoudhury, Shin Hwei Tan, Patanamon Thongtanunam
Abstract: Automatic programming has seen increasing popularity due to the emergence of tools like GitHub Copilot which rely on Large Language Models (LLMs). At the same time, automatically generated code faces challenges during deployment due to concerns around quality and trust. In this article, we study automated coding in a general sense and study the concerns around code quality, security and related issues of programmer responsibility. These are key issues for organizations while deciding on the usage of automatically generated code. We discuss how advances in software engineering such as program repair and analysis can enable automatic programming. We conclude with a forward looking view, focusing on the programming environment of the near future, where programmers may need to switch to different roles to fully utilize the power of automatic programming. Automated repair of automatically generated programs from LLMs, can help produce higher assurance code from LLMs, along with evidence of assurance
Authors: Brayan Monroy, Juan Estupi\~nan, Tatiana Gelvez-Barrera, Jorge Bacca, Henry Arguello
Abstract: Binary Neural Networks emerged as a cost-effective and energy-efficient solution for computer vision tasks by binarizing either network weights or activations. However, common binary activations, such as the Sign activation function, abruptly binarize the values with a single threshold, losing fine-grained details in the feature outputs. This work proposes an activation that applies multiple thresholds following dithering principles, shifting the Sign activation function for each pixel according to a spatially periodic threshold kernel. Unlike literature methods, the shifting is defined jointly for a set of adjacent pixels, taking advantage of spatial correlations. Experiments over the classification task demonstrate the effectiveness of the designed dithering Sign activation function as an alternative activation for binary neural networks, without increasing the computational cost. Further, DeSign balances the preservation of details with the efficiency of binary operations.
Authors: Samuel Lanthaler, Andrew M. Stuart, Margaret Trautner
Abstract: Operator learning is a variant of machine learning that is designed to approximate maps between function spaces from data. The Fourier Neural Operator (FNO) is a common model architecture used for operator learning. The FNO combines pointwise linear and nonlinear operations in physical space with pointwise linear operations in Fourier space, leading to a parameterized map acting between function spaces. Although FNOs formally involve convolutions of functions on a continuum, in practice the computations are performed on a discretized grid, allowing efficient implementation via the FFT. In this paper, the aliasing error that results from such a discretization is quantified and algebraic rates of convergence in terms of the grid resolution are obtained as a function of the regularity of the input. Numerical experiments that validate the theory and describe model stability are performed.
Authors: Lujing Zhang, Aaron Roth, Linjun Zhang
Abstract: This paper introduces a framework for post-processing machine learning models so that their predictions satisfy multi-group fairness guarantees. Based on the celebrated notion of multicalibration, we introduce $(\mathbf{s},\mathcal{G}, \alpha)-$GMC (Generalized Multi-Dimensional Multicalibration) for multi-dimensional mappings $\mathbf{s}$, constraint set $\mathcal{G}$, and a pre-specified threshold level $\alpha$. We propose associated algorithms to achieve this notion in general settings. This framework is then applied to diverse scenarios encompassing different fairness concerns, including false negative rate control in image segmentation, prediction set conditional uncertainty quantification in hierarchical classification, and de-biased text generation in language models. We conduct numerical studies on several datasets and tasks.
Authors: Xin Wang, Ziwei Zhang, Haoyang Li, Wenwu Zhu
Abstract: Graph machine learning has been extensively studied in both academic and industry. However, as the literature on graph learning booms with a vast number of emerging methods and techniques, it becomes increasingly difficult to manually design the optimal machine learning algorithm for different graph-related tasks. To tackle the challenge, automated graph machine learning, which aims at discovering the best hyper-parameter and neural architecture configuration for different graph tasks/data without manual design, is gaining an increasing number of attentions from the research community. In this paper, we extensively discuss automated graph machine learning approaches, covering hyper-parameter optimization (HPO) and neural architecture search (NAS) for graph machine learning. We briefly overview existing libraries designed for either graph machine learning or automated machine learning respectively, and further in depth introduce AutoGL, our dedicated and the world's first open-source library for automated graph machine learning. Also, we describe a tailored benchmark that supports unified, reproducible, and efficient evaluations. Last but not least, we share our insights on future research directions for automated graph machine learning. This paper is the first systematic and comprehensive discussion of approaches, libraries as well as directions for automated graph machine learning.
Authors: Silvia Beddar-Wiesing, Giuseppe Alessio D'Inverno, Caterina Graziani, Veronica Lachi, Alice Moallemy-Oureh, Franco Scarselli, Josephine Maria Thomas
Abstract: Graph Neural Networks (GNNs) are a large class of relational models for graph processing. Recent theoretical studies on the expressive power of GNNs have focused on two issues. On the one hand, it has been proven that GNNs are as powerful as the Weisfeiler-Lehman test (1-WL) in their ability to distinguish graphs. Moreover, it has been shown that the equivalence enforced by 1-WL equals unfolding equivalence. On the other hand, GNNs turned out to be universal approximators on graphs modulo the constraints enforced by 1-WL/unfolding equivalence. However, these results only apply to Static Attributed Undirected Homogeneous Graphs (SAUHG) with node attributes. In contrast, real-life applications often involve a much larger variety of graph types. In this paper, we conduct a theoretical analysis of the expressive power of GNNs for two other graph domains that are particularly interesting in practical applications, namely dynamic graphs and SAUGHs with edge attributes. Dynamic graphs are widely used in modern applications; hence, the study of the expressive capability of GNNs in this domain is essential for practical reasons and, in addition, it requires a new analyzing approach due to the difference in the architecture of dynamic GNNs compared to static ones. On the other hand, the examination of SAUHGs is of particular relevance since they act as a standard form for all graph types: it has been shown that all graph types can be transformed without loss of information to SAUHGs with both attributes on nodes and edges. This paper considers generic GNN models and appropriate 1-WL tests for those domains. Then, the known results on the expressive power of GNNs are extended to the mentioned domains: it is proven that GNNs have the same capability as the 1-WL test, the 1-WL equivalence equals unfolding equivalence and that GNNs are universal approximators modulo 1-WL/unfolding equivalence.
Authors: Ching Chang, Wen-Chih Peng
Abstract: With the rapid development of technology, the automated monitoring systems of large-scale factories are becoming more and more important. By collecting a large amount of machine sensor data, we can have many ways to find anomalies. We believe that the real core value of an automated monitoring system is to identify and track the cause of the problem. The most famous method for finding causal anomalies is RCA, but there are many problems that cannot be ignored. They used the AutoRegressive eXogenous (ARX) model to create a time-invariant correlation network as a machine profile, and then use this profile to track the causal anomalies by means of a method called fault propagation. There are two major problems in describing the behavior of a machine by using the correlation network established by ARX: (1) It does not take into account the diversity of states (2) It does not separately consider the correlations with different time-lag. Based on these problems, we propose a framework called Ranking Causal Anomalies in End-to-End System (RCAE2E), which completely solves the problems mentioned above. In the experimental part, we use synthetic data and real-world large-scale photoelectric factory data to verify the correctness and existence of our method hypothesis.
Authors: Juyoung Yun, Byungkon Kang, Zhoulai Fu
Abstract: Lowering the precision of neural networks from the prevalent 32-bit precision has long been considered harmful to performance, despite the gain in space and time. Many works propose various techniques to implement half-precision neural networks, but none study pure 16-bit settings. This paper investigates the unexpected performance gain of pure 16-bit neural networks over the 32-bit networks in classification tasks. We present extensive experimental results that favorably compare various 16-bit neural networks' performance to those of the 32-bit models. In addition, a theoretical analysis of the efficiency of 16-bit models is provided, which is coupled with empirical evidence to back it up. Finally, we discuss situations in which low-precision training is indeed detrimental.
Authors: Xiyuan Wang, Haotong Yang, Muhan Zhang
Abstract: In this work, we propose a novel link prediction model and further boost it by studying graph incompleteness. First, we introduce MPNN-then-SF, an innovative architecture leveraging structural feature (SF) to guide MPNN's representation pooling, with its implementation, namely Neural Common Neighbor (NCN). NCN exhibits superior expressiveness and scalability compared with existing models, which can be classified into two categories: SF-then-MPNN, augmenting MPNN's input with SF, and SF-and-MPNN, decoupling SF and MPNN. Second, we investigate the impact of graph incompleteness -- the phenomenon that some links are unobserved in the input graph -- on SF, like the common neighbor. Through dataset visualization, we observe that incompleteness reduces common neighbors and induces distribution shifts, significantly affecting model performance. To address this issue, we propose to use a link prediction model to complete the common neighbor structure. Combining this method with NCN, we propose Neural Common Neighbor with Completion (NCNC). NCN and NCNC outperform recent strong baselines by large margins, and NCNC further surpasses state-of-the-art models in standard link prediction benchmarks. Our code is available at https://github.com/GraphPKU/NeuralCommonNeighbor.
Authors: Hui-Po Wang, Dingfan Chen, Raouf Kerkouche, Mario Fritz
Abstract: Conventional gradient-sharing approaches for federated learning (FL), such as FedAvg, rely on aggregation of local models and often face performance degradation under differential privacy (DP) mechanisms or data heterogeneity, which can be attributed to the inconsistency between the local and global objectives. To address this issue, we propose FedLAP-DP, a novel privacy-preserving approach for FL. Our formulation involves clients synthesizing a small set of samples that approximate local loss landscapes by simulating the gradients of real images within a local region. Acting as loss surrogates, these synthetic samples are aggregated on the server side to uncover the global loss landscape and enable global optimization. Building upon these insights, we offer a new perspective to enforce record-level differential privacy in FL. A formal privacy analysis demonstrates that FedLAP-DP incurs the same privacy costs as typical gradient-sharing schemes while achieving an improved trade-off between privacy and utility. Extensive experiments validate the superiority of our approach across various datasets with highly skewed distributions in both DP and non-DP settings. Beyond the promising performance, our approach presents a faster convergence speed compared to typical gradient-sharing methods and opens up the possibility of trading communication costs for better performance by sending a larger set of synthetic images. The source is available at https://github.com/a514514772/FedLAP-DP.
Authors: Yong Qing, Ke Li, Peng-Fei Zhou, Shi-Ju Ran
Abstract: Neural network (NN) designed for challenging machine learning tasks is in general a highly nonlinear mapping that contains massive variational parameters. High complexity of NN, if unbounded or unconstrained, might unpredictably cause severe issues including over-fitting, loss of generalization power, and unbearable cost of hardware. In this work, we propose a general compression scheme that significantly reduces the variational parameters of NN by encoding them to deep automatically-differentiable tensor network (ADTN) that contains exponentially-fewer free parameters. Superior compression performance of our scheme is demonstrated on several widely-recognized NN's (FC-2, LeNet-5, AlextNet, ZFNet and VGG-16) and datasets (MNIST, CIFAR-10 and CIFAR-100). For instance, we compress two linear layers in VGG-16 with approximately $10^{7}$ parameters to two ADTN's with just 424 parameters, where the testing accuracy on CIFAR-10 is improved from $90.17 \%$ to $91.74\%$. Our work suggests TN as an exceptionally efficient mathematical structure for representing the variational parameters of NN's, which exhibits superior compressibility over the commonly-used matrices and multi-way arrays.
Authors: Anna-Christina Glock, Florian Sobieczky, Johannes F\"urnkranz, Peter Filzmoser, Martin Jech
Abstract: A change point detection (CPD) framework assisted by a predictive machine learning model called "Predict and Compare" is introduced and characterised in relation to other state-of-the-art online CPD routines which it outperforms in terms of false positive rate and out-of-control average run length. The method's focus is on improving standard methods from sequential analysis such as the CUSUM rule in terms of these quality measures. This is achieved by replacing typically used trend estimation functionals such as the running mean with more sophisticated predictive models (Predict step), and comparing their prognosis with actual data (Compare step). The two models used in the Predict step are the ARIMA model and the LSTM recursive neural network. However, the framework is formulated in general terms, so as to allow the use of other prediction or comparison methods than those tested here. The power of the method is demonstrated in a tribological case study in which change points separating the run-in, steady-state, and divergent wear phases are detected in the regime of very few false positives.
Authors: Haozhe Jiang, Qiwen Cui, Zhihan Xiong, Maryam Fazel, Simon S. Du
Abstract: We investigate learning the equilibria in non-stationary multi-agent systems and address the challenges that differentiate multi-agent learning from single-agent learning. Specifically, we focus on games with bandit feedback, where testing an equilibrium can result in substantial regret even when the gap to be tested is small, and the existence of multiple optimal solutions (equilibria) in stationary games poses extra challenges. To overcome these obstacles, we propose a versatile black-box approach applicable to a broad spectrum of problems, such as general-sum games, potential games, and Markov games, when equipped with appropriate learning and testing oracles for stationary environments. Our algorithms can achieve $\widetilde{O}\left(\Delta^{1/4}T^{3/4}\right)$ regret when the degree of nonstationarity, as measured by total variation $\Delta$, is known, and $\widetilde{O}\left(\Delta^{1/5}T^{4/5}\right)$ regret when $\Delta$ is unknown, where $T$ is the number of rounds. Meanwhile, our algorithm inherits the favorable dependence on number of agents from the oracles. As a side contribution that may be independent of interest, we show how to test for various types of equilibria by a black-box reduction to single-agent learning, which includes Nash equilibria, correlated equilibria, and coarse correlated equilibria.
Authors: Joern Ploennigs, Markus Berger
Abstract: AI image generators based on diffusion models have recently garnered attention for their capability to create images from simple text prompts. However, for practical use in civil engineering they need to be able to create specific construction plans for given constraints. This paper investigates the potential of current AI generators in addressing such challenges, specifically for the creation of simple floor plans. We explain how the underlying diffusion-models work and propose novel refinement approaches to improve semantic encoding and generation quality. In several experiments we show that we can improve validity of generated floor plans from 6% to 90%. Based on these results we derive future research challenges considering building information modelling. With this we provide: (i) evaluation of current generative AIs; (ii) propose improved refinement approaches; (iii) evaluate them on various examples; (iv) derive future directions for diffusion models in civil engineering.
Authors: Laura Iacovissi, Nan Lu, Robert C. Williamson
Abstract: Corruption is notoriously widespread in data collection. Despite extensive research, the existing literature on corruption predominantly focuses on specific settings and learning scenarios, lacking a unified view. There is still a limited understanding of how to effectively model and mitigate corruption in machine learning problems. In this work, we develop a general theory of corruption from an information-theoretic perspective - with Markov kernels as a foundational mathematical tool. We generalize the definition of corruption beyond the concept of distributional shift: corruption includes all modifications of a learning problem, including changes in model class and loss function. We will focus here on changes in probability distributions. First, we construct a provably exhaustive framework for pairwise Markovian corruptions. The framework not only allows us to study corruption types based on their input space, but also serves to unify prior works on specific corruption models and establish a consistent nomenclature. Second, we systematically analyze the consequences of corruption on learning tasks by comparing Bayes risks in the clean and corrupted scenarios. This examination sheds light on complexities arising from joint and dependent corruptions on both labels and attributes. Notably, while label corruptions affect only the loss function, more intricate cases involving attribute corruptions extend the influence beyond the loss to affect the hypothesis class. Third, building upon these results, we investigate mitigations for various corruption types. We expand the existing loss-correction results for label corruption, and identify the necessity to generalize the classical corruption-corrected learning framework to a new paradigm with weaker requirements. Within the latter setting, we provide a negative result for loss correction in the attribute and the joint corruption case.
Authors: Yi Cai, Gerhard Wunder
Abstract: Attribution methods shed light on the explainability of data-driven approaches such as deep learning models by uncovering the most influential features in a to-be-explained decision. While determining feature attributions via gradients delivers promising results, the internal access required for acquiring gradients can be impractical under safety concerns, thus limiting the applicability of gradient-based approaches. In response to such limited flexibility, this paper presents \methodAbr~(gradient-estimation-based explanation), an approach that produces gradient-like explanations through only query-level access. The proposed approach holds a set of fundamental properties for attribution methods, which are mathematically rigorously proved, ensuring the quality of its explanations. In addition to the theoretical analysis, with a focus on image data, the experimental results empirically demonstrate the superiority of the proposed method over state-of-the-art black-box methods and its competitive performance compared to methods with full access.
Authors: Zijun Lin, Haidi Azaman, M Ganesh Kumar, Cheston Tan
Abstract: Children can rapidly generalize compositionally-constructed rules to unseen test sets. On the other hand, deep reinforcement learning (RL) agents need to be trained over millions of episodes, and their ability to generalize to unseen combinations remains unclear. Hence, we investigate the compositional abilities of RL agents, using the task of navigating to specified color-shape targets in synthetic 3D environments. First, we show that when RL agents are naively trained to navigate to target color-shape combinations, they implicitly learn to decompose the combinations, allowing them to (re-)compose these and succeed at held-out test combinations ("compositional learning"). Second, when agents are pretrained to learn invariant shape and color concepts ("concept learning"), the number of episodes subsequently needed for compositional learning decreased by 20 times. Furthermore, only agents trained on both concept and compositional learning could solve a more complex, out-of-distribution environment in zero-shot fashion. Finally, we verified that only text encoders pretrained on image-text datasets (e.g. CLIP) reduced the number of training episodes needed for our agents to demonstrate compositional learning, and also generalized to 5 unseen colors in zero-shot fashion. Overall, our results are the first to demonstrate that RL agents can be trained to implicitly learn concepts and compositionality, to solve more complex environments in zero-shot fashion.
Authors: Andrew F. Luo, Margaret M. Henderson, Michael J. Tarr, Leila Wehbe
Abstract: Understanding the functional organization of higher visual cortex is a central focus in neuroscience. Past studies have primarily mapped the visual and semantic selectivity of neural populations using hand-selected stimuli, which may potentially bias results towards pre-existing hypotheses of visual cortex functionality. Moving beyond conventional approaches, we introduce a data-driven method that generates natural language descriptions for images predicted to maximally activate individual voxels of interest. Our method -- Semantic Captioning Using Brain Alignments ("BrainSCUBA") -- builds upon the rich embedding space learned by a contrastive vision-language model and utilizes a pre-trained large language model to generate interpretable captions. We validate our method through fine-grained voxel-level captioning across higher-order visual regions. We further perform text-conditioned image synthesis with the captions, and show that our images are semantically coherent and yield high predicted activations. Finally, to demonstrate how our method enables scientific discovery, we perform exploratory investigations on the distribution of "person" representations in the brain, and discover fine-grained semantic selectivity in body-selective areas. Unlike earlier studies that decode text, our method derives voxel-wise captions of semantic selectivity. Our results show that BrainSCUBA is a promising means for understanding functional preferences in the brain, and provides motivation for further hypothesis-driven investigation of visual cortex.
Authors: Samet Bayram, Kenneth Barner
Abstract: This paper presents GReAT (Graph Regularized Adversarial Training), a novel regularization method designed to enhance the robust classification performance of deep learning models. Adversarial examples, characterized by subtle perturbations that can mislead models, pose a significant challenge in machine learning. Although adversarial training is effective in defending against such attacks, it often overlooks the underlying data structure. In response, GReAT integrates graph based regularization into the adversarial training process, leveraging the data's inherent structure to enhance model robustness. By incorporating graph information during training, GReAT defends against adversarial attacks and improves generalization to unseen data. Extensive evaluations on benchmark datasets demonstrate that GReAT outperforms state of the art methods in robustness, achieving notable improvements in classification accuracy. Specifically, compared to the second best methods, GReAT achieves a performance increase of approximately 4.87% for CIFAR10 against FGSM attack and 10.57% for SVHN against FGSM attack. Additionally, for CIFAR10, GReAT demonstrates a performance increase of approximately 11.05% against PGD attack, and for SVHN, a 5.54% increase against PGD attack. This paper provides detailed insights into the proposed methodology, including numerical results and comparisons with existing approaches, highlighting the significant impact of GReAT in advancing the performance of deep learning models.
Authors: Tim De Ryck, Florent Bonnet, Siddhartha Mishra, Emmanuel de B\'ezenac
Abstract: In this paper, we investigate the behavior of gradient descent algorithms in physics-informed machine learning methods like PINNs, which minimize residuals connected to partial differential equations (PDEs). Our key result is that the difficulty in training these models is closely related to the conditioning of a specific differential operator. This operator, in turn, is associated to the Hermitian square of the differential operator of the underlying PDE. If this operator is ill-conditioned, it results in slow or infeasible training. Therefore, preconditioning this operator is crucial. We employ both rigorous mathematical analysis and empirical evaluations to investigate various strategies, explaining how they better condition this critical operator, and consequently improve training.
Authors: Theo Archambault, Arthur Filoche, Anastase Charantonis, Dominique Bereziat, Sylvie Thiria
Abstract: Satellite-based remote sensing missions have revolutionized our understanding of the Ocean state and dynamics. Among them, space-borne altimetry provides valuable measurements of Sea Surface Height (SSH), which is used to estimate surface geostrophic currents. Due to the sensor technology employed, important gaps occur in SSH observations. Complete SSH maps are produced using linear Optimal Interpolations (OI) such as the widely-used Data Unification and Altimeter Combination System (duacs). On the other hand, Sea Surface Temperature (SST) products have much higher data coverage and SST is physically linked to geostrophic currents through advection. We propose a new multi-variate Observing System Simulation Experiment (OSSE) emulating 20 years of SSH and SST satellite observations. We train an Attention-Based Encoder-Decoder deep learning network (abed) on this data, comparing two settings: one with access to ground truth during training and one without. On our OSSE, we compare abed reconstructions when trained using either supervised or unsupervised loss functions, with or without SST information. We evaluate the SSH interpolations in terms of eddy detection. We also introduce a new way to transfer the learning from simulation to observations by doing a supervised pre-training on our OSSE followed by an unsupervised fine-tuning on satellite data. On real SSH observations from the Ocean Data Challenge 2021, we find that this learning strategy combined with the use of SST leads to a decrease of 24% of the root mean squared error compared to duacs.
Authors: Fan Wu, Huseyin A. Inan, Arturs Backurs, Varun Chandrasekaran, Janardhan Kulkarni, Robert Sim
Abstract: Positioned between pre-training and user deployment, aligning large language models (LLMs) through reinforcement learning (RL) has emerged as a prevailing strategy for training instruction following-models such as ChatGPT. In this work, we initiate the study of privacy-preserving alignment of LLMs through Differential Privacy (DP) in conjunction with RL. Following the influential work of Ziegler et al. (2020), we study two dominant paradigms: (i) alignment via RL without human in the loop (e.g., positive review generation) and (ii) alignment via RL from human feedback (RLHF) (e.g., summarization in a human-preferred way). We give a new DP framework to achieve alignment via RL, and prove its correctness. Our experimental results validate the effectiveness of our approach, offering competitive utility while ensuring strong privacy protections.
Authors: Aaron Archer, Matthew Fahrbach, Kuikui Liu, Prakash Prabhu
Abstract: We optimize pipeline parallelism for deep neural network (DNN) inference by partitioning model graphs into $k$ stages and minimizing the running time of the bottleneck stage, including communication. We give practical and effective algorithms for this NP-hard problem, but our emphasis is on tackling the practitioner's dilemma of deciding when a solution is good enough. To this end, we design novel mixed-integer programming (MIP) relaxations for proving lower bounds. Applying these methods to a diverse testbed of 369 production models, for $k \in \{2, 4, 8, 16, 32, 64\}$, we empirically show that these lower bounds are strong enough to be useful in practice. Our lower bounds are substantially stronger than standard combinatorial bounds. For example, evaluated via geometric means across our production testbed with $k = 16$ pipeline stages, our MIP formulations raised the lower bound from 0.4598 to 0.9452, expressed as a fraction of the best partition found. In other words, our improved lower bounds closed the optimality gap by a factor of 9.855x.
Authors: Joohyung Lee, Mohamed Seif, Jungchan Cho, H. Vincent Poor
Abstract: Split Federated Learning (SFL) has recently emerged as a promising distributed learning technology, leveraging the strengths of both federated and split learning. It emphasizes the advantages of rapid convergence while addressing privacy concerns. As a result, this innovation has received significant attention from both industry and academia. However, since the model is split at a specific layer, known as a cut layer, into both client-side and server-side models for the SFL, the choice of the cut layer in SFL can have a substantial impact on the energy consumption of clients and their privacy, as it influences the training burden and the output of the client-side models. In this article, we provide a comprehensive overview of the SFL process and thoroughly analyze energy consumption and privacy. This analysis considers the influence of various system parameters on the cut layer selection strategy. Additionally, we provide an illustrative example of the cut layer selection, aiming to minimize clients' risk of reconstructing the raw data at the server while sustaining energy consumption within the required energy budget, which involves trade-offs. Finally, we address open challenges in this field. These directions represent promising avenues for future research and development.
Authors: Qiqi Su, Christos Kloukinas, Artur d'Avila Garcez
Abstract: Multivariate time series have many applications, from healthcare and meteorology to life science. Although deep learning models have shown excellent predictive performance for time series, they have been criticised for being "black-boxes" or non-interpretable. This paper proposes a novel modular neural network model for multivariate time series prediction that is interpretable by construction. A recurrent neural network learns the temporal dependencies in the data while an attention-based feature selection component selects the most relevant features and suppresses redundant features used in the learning of the temporal dependencies. A modular deep network is trained from the selected features independently to show the users how features influence outcomes, making the model interpretable. Experimental results show that this approach can outperform state-of-the-art interpretable Neural Additive Models (NAM) and variations thereof in both regression and classification of time series tasks, achieving a predictive performance that is comparable to the top non-interpretable methods for time series, LSTM and XGBoost.
Authors: Jiayuan Mao, Tom\'as Lozano-P\'erez, Joshua B. Tenenbaum, Leslie Pack Kaelbling
Abstract: Goal-conditioned policies are generally understood to be "feed-forward" circuits, in the form of neural networks that map from the current state and the goal specification to the next action to take. However, under what circumstances such a policy can be learned and how efficient the policy will be are not well understood. In this paper, we present a circuit complexity analysis for relational neural networks (such as graph neural networks and transformers) representing policies for planning problems, by drawing connections with serialized goal regression search (S-GRS). We show that there are three general classes of planning problems, in terms of the growth of circuit width and depth as a function of the number of objects and planning horizon, providing constructive proofs. We also illustrate the utility of this analysis for designing neural networks for policy learning.
Authors: Nil-Jana Akpinar, Zachary C. Lipton, Alexandra Chouldechova
Abstract: Predictive risk models in the public sector are commonly developed using administrative data that is more complete for subpopulations that more greatly rely on public services. In the United States, for instance, information on health care utilization is routinely available to government agencies for individuals supported by Medicaid and Medicare, but not for the privately insured. Critiques of public sector algorithms have identified such differential feature under-reporting as a driver of disparities in algorithmic decision-making. Yet this form of data bias remains understudied from a technical viewpoint. While prior work has examined the fairness impacts of additive feature noise and features that are clearly marked as missing, the setting of data missingness absent indicators (i.e. differential feature under-reporting) has been lacking in research attention. In this work, we present an analytically tractable model of differential feature under-reporting which we then use to characterize the impact of this kind of data bias on algorithmic fairness. We demonstrate how standard missing data methods typically fail to mitigate bias in this setting, and propose a new set of methods specifically tailored to differential feature under-reporting. Our results show that, in real world data settings, under-reporting typically leads to increasing disparities. The proposed solution methods show success in mitigating increases in unfairness.
Authors: Michele Caprio, Maryam Sultana, Eleni Elia, Fabio Cuzzolin
Abstract: Statistical learning theory is the foundation of machine learning, providing theoretical bounds for the risk of models learnt from a (single) training set, assumed to issue from an unknown probability distribution. In actual deployment, however, the data distribution may (and often does) vary, causing domain adaptation/generalization issues. In this paper we lay the foundations for a `credal' theory of learning, using convex sets of probabilities (credal sets) to model the variability in the data-generating distribution. Such credal sets, we argue, may be inferred from a finite sample of training sets. Bounds are derived for the case of finite hypotheses spaces (both assuming realizability or not) as well as infinite model spaces, which directly generalize classical results.
Authors: Buyun Zhang, Liang Luo, Yuxin Chen, Jade Nie, Xi Liu, Daifeng Guo, Yanli Zhao, Shen Li, Yuchen Hao, Yantao Yao, Guna Lakshminarayanan, Ellie Dingqiao Wen, Jongsoo Park, Maxim Naumov, Wenlin Chen
Abstract: Scaling laws play an instrumental role in the sustainable improvement in model quality. Unfortunately, recommendation models to date do not exhibit such laws similar to those observed in the domain of large language models, due to the inefficiencies of their upscaling mechanisms. This limitation poses significant challenges in adapting these models to increasingly more complex real-world datasets. In this paper, we propose an effective network architecture based purely on stacked factorization machines, and a synergistic upscaling strategy, collectively dubbed Wukong, to establish a scaling law in the domain of recommendation. Wukong's unique design makes it possible to capture diverse, any-order of interactions simply through taller and wider layers. We conducted extensive evaluations on six public datasets, and our results demonstrate that Wukong consistently outperforms state-of-the-art models quality-wise. Further, we assessed Wukong's scalability on an internal, large-scale dataset. The results show that Wukong retains its superiority in quality over state-of-the-art models, while holding the scaling law across two orders of magnitude in model complexity, extending beyond 100 Gflop or equivalently up to Large Language Model (GPT-3) training compute scale, where prior arts fall short.
Authors: Max Rudolph, Caleb Chuck, Kevin Black, Misha Lvovsky, Scott Niekum, Amy Zhang
Abstract: Robust reinforcement learning agents using high-dimensional observations must be able to identify relevant state features amidst many exogeneous distractors. A representation that captures controllability identifies these state elements by determining what affects agent control. While methods such as inverse dynamics and mutual information capture controllability for a limited number of timesteps, capturing long-horizon elements remains a challenging problem. Myopic controllability can capture the moment right before an agent crashes into a wall, but not the control-relevance of the wall while the agent is still some distance away. To address this we introduce action-bisimulation encoding, a method inspired by the bisimulation invariance pseudometric, that extends single-step controllability with a recursive invariance constraint. By doing this, action-bisimulation learns a multi-step controllability metric that smoothly discounts distant state features that are relevant for control. We demonstrate that action-bisimulation pretraining on reward-free, uniformly random data improves sample efficiency in several environments, including a photorealistic 3D simulation domain, Habitat. Additionally, we provide theoretical analysis and qualitative results demonstrating the information captured by action-bisimulation.
Authors: Zhenjiang Mao, Siqi Dai, Yuang Geng, Ivan Ruchkin
Abstract: A world model creates a surrogate world to train a controller and predict safety violations by learning the internal dynamic model of systems. However, the existing world models rely solely on statistical learning of how observations change in response to actions, lacking precise quantification of how accurate the surrogate dynamics are, which poses a significant challenge in safety-critical systems. To address this challenge, we propose foundation world models that embed observations into meaningful and causally latent representations. This enables the surrogate dynamics to directly predict causal future states by leveraging a training-free large language model. In two common benchmarks, this novel model outperforms standard world models in the safety prediction task and has a performance comparable to supervised learning despite not using any data. We evaluate its performance with a more specialized and system-relevant metric by comparing estimated states instead of aggregating observation-wide error.
Authors: Sean Farhat, Deming Chen
Abstract: In this paper, we propose that small models may not need to absorb the cost of pre-training to reap its benefits. Instead, they can capitalize on the astonishing results achieved by modern, enormous models to a surprising degree. We observe that, when distilled on a task from a pre-trained teacher model, a small model can achieve or surpass the performance it would achieve if it was pre-trained then finetuned on that task. To allow this phenomenon to be easily leveraged, we establish a connection reducing knowledge distillation to modern contrastive learning, opening two doors: (1) vastly different model architecture pairings can work for the distillation, and (2) most contrastive learning algorithms rooted in the theory of Noise Contrastive Estimation can be easily applied and used. We demonstrate this paradigm using pre-trained teacher models from open-source model hubs, Transformer and convolution based model combinations, and a novel distillation algorithm that massages the Alignment/Uniformity perspective of contrastive learning by Wang & Isola (2020) into a distillation objective. We choose this flavor of contrastive learning due to its low computational cost, an overarching theme of this work. We also observe that this phenomenon tends not to occur if the task is data-limited. However, this can be alleviated by leveraging yet another scale-inspired development: large, pre-trained generative models for dataset augmentation. Again, we use an open-source model, and our rudimentary prompts are sufficient to boost the small model`s performance. Thus, we highlight a training method for small models that is up to 94% faster than the standard pre-training paradigm without sacrificing performance. For practitioners discouraged from fully utilizing modern foundation datasets for their small models due to the prohibitive scale, we believe our work keeps that door open.
Authors: Miguel Costa, Sandro Pinto
Abstract: ML is shifting from the cloud to the edge. Edge computing reduces the surface exposing private data and enables reliable throughput guarantees in real-time applications. Of the panoply of devices deployed at the edge, resource-constrained MCUs, e.g., Arm Cortex-M, are more prevalent, orders of magnitude cheaper, and less power-hungry than application processors or GPUs. Thus, enabling intelligence at the deep edge is the zeitgeist, with researchers focusing on unveiling novel approaches to deploy ANNs on these constrained devices. Quantization is a well-established technique that has proved effective in enabling the deployment of neural networks on MCUs; however, it is still an open question to understand the robustness of QNNs in the face of adversarial examples. To fill this gap, we empirically evaluate the effectiveness of attacks and defenses from (full-precision) ANNs on (constrained) QNNs. Our evaluation includes three QNNs targeting TinyML applications, ten attacks, and six defenses. With this study, we draw a set of interesting findings. First, quantization increases the point distance to the decision boundary and leads the gradient estimated by some attacks to explode or vanish. Second, quantization can act as a noise attenuator or amplifier, depending on the noise magnitude, and causes gradient misalignment. Regarding adversarial defenses, we conclude that input pre-processing defenses show impressive results on small perturbations; however, they fall short as the perturbation increases. At the same time, train-based defenses increase the average point distance to the decision boundary, which holds after quantization. However, we argue that train-based defenses still need to smooth the quantization-shift and gradient misalignment phenomenons to counteract adversarial example transferability to QNNs. All artifacts are open-sourced to enable independent validation of results.
Authors: Andre R Kuroswiski, Annie S Wu, Angelo Passaro
Abstract: In this paper, we introduce an alternative approach to enhancing Multi-Agent Reinforcement Learning (MARL) through the integration of domain knowledge and attention-based policy mechanisms. Our methodology focuses on the incorporation of domain-specific expertise into the learning process, which simplifies the development of collaborative behaviors. This approach aims to reduce the complexity and learning overhead typically associated with MARL by enabling agents to concentrate on essential aspects of complex tasks, thus optimizing the learning curve. The utilization of attention mechanisms plays a key role in our model. It allows for the effective processing of dynamic context data and nuanced agent interactions, leading to more refined decision-making. Applied in standard MARL scenarios, such as the Stanford Intelligent Systems Laboratory (SISL) Pursuit and Multi-Particle Environments (MPE) Simple Spread, our method has been shown to improve both learning efficiency and the effectiveness of collaborative behaviors. The results indicate that our attention-based approach can be a viable approach for improving the efficiency of MARL training process, integrating domain-specific knowledge at the action level.
Authors: Doron Haviv, Russell Zhang Kunes, Thomas Dougherty, Cassandra Burdziak, Tal Nawy, Anna Gilbert, Dana Pe'er
Abstract: Optimal transport (OT) and the related Wasserstein metric (W) are powerful and ubiquitous tools for comparing distributions. However, computing pairwise Wasserstein distances rapidly becomes intractable as cohort size grows. An attractive alternative would be to find an embedding space in which pairwise Euclidean distances map to OT distances, akin to standard multidimensional scaling (MDS). We present Wasserstein Wormhole, a transformer-based autoencoder that embeds empirical distributions into a latent space wherein Euclidean distances approximate OT distances. Extending MDS theory, we show that our objective function implies a bound on the error incurred when embedding non-Euclidean distances. Empirically, distances between Wormhole embeddings closely match Wasserstein distances, enabling linear time computation of OT distances. Along with an encoder that maps distributions to embeddings, Wasserstein Wormhole includes a decoder that maps embeddings back to distributions, allowing for operations in the embedding space to generalize to OT spaces, such as Wasserstein barycenter estimation and OT interpolation. By lending scalability and interpretability to OT approaches, Wasserstein Wormhole unlocks new avenues for data analysis in the fields of computational geometry and single-cell biology.
Authors: Sibo Gai, Donglin Wang
Abstract: In this paper, we study the continual learning problem of single-task offline reinforcement learning. In the past, continual reinforcement learning usually only dealt with multitasking, that is, learning multiple related or unrelated tasks in a row, but once each learned task was learned, it was not relearned, but only used in subsequent processes. However, offline reinforcement learning tasks require the continuously learning of multiple different datasets for the same task. Existing algorithms will try their best to achieve the best results in each offline dataset they have learned and the skills of the network will overwrite the high-quality datasets that have been learned after learning the subsequent poor datasets. On the other hand, if too much emphasis is placed on stability, the network will learn the subsequent better dataset after learning the poor offline dataset, and the problem of insufficient plasticity and non-learning will occur. How to design a strategy that can always preserve the best performance for each state in the data that has been learned is a new challenge and the focus of this study. Therefore, this study proposes a new algorithm, called Ensemble Offline Reinforcement Learning Based on Experience Replay, which introduces multiple value networks to learn the same dataset and judge whether the strategy has been learned by the discrete degree of the value network, to improve the performance of the network in single-task offline reinforcement learning.
Authors: Tahrima Hashem, Negin Yousefpour
Abstract: Scour around bridge piers is a critical challenge for infrastructures around the world. In the absence of analytical models and due to the complexity of the scour process, it is difficult for current empirical methods to achieve accurate predictions. In this paper, we exploit the power of deep learning algorithms to forecast the scour depth variations around bridge piers based on historical sensor monitoring data, including riverbed elevation, flow elevation, and flow velocity. We investigated the performance of Long Short-Term Memory (LSTM) and Convolutional Neural Network (CNN) models for real-time scour forecasting using data collected from bridges in Alaska and Oregon from 2006 to 2021. The LSTM models achieved mean absolute error (MAE) ranging from 0.1m to 0.5m for predicting bed level variations a week in advance, showing a reasonable performance. The Fully Convolutional Network (FCN) variant of CNN outperformed other CNN configurations, showing a comparable performance to LSTMs with significantly lower computational costs. We explored various innovative random-search heuristics for hyperparameter tuning and model optimisation which resulted in reduced computational cost compared to grid-search method. The impact of different combinations of sensor features on scour prediction showed the significance of the historical time series of scour for predicting upcoming events. Overall, this study provides a greater understanding of the potential of Deep Learning algorithms for real-time scour prediction and early warning for bridges with distinct geology, geomorphology and flow characteristics.
Authors: Francis Ogoke, Peter Myung-Won Pak, Alexander Myers, Guadalupe Quirarte, Jack Beuth, Jonathan Malen, Amir Barati Farimani
Abstract: Insufficient overlap between the melt pools produced during Laser Powder Bed Fusion (L-PBF) can lead to lack-of-fusion defects and deteriorated mechanical and fatigue performance. In-situ monitoring of the melt pool subsurface morphology requires specialized equipment that may not be readily accessible or scalable. Therefore, we introduce a machine learning framework to correlate in-situ two-color thermal images observed via high-speed color imaging to the two-dimensional profile of the melt pool cross-section. Specifically, we employ a hybrid CNN-Transformer architecture to establish a correlation between single bead off-axis thermal image sequences and melt pool cross-section contours measured via optical microscopy. In this architecture, a ResNet model embeds the spatial information contained within the thermal images to a latent vector, while a Transformer model correlates the sequence of embedded vectors to extract temporal information. Our framework is able to model the curvature of the subsurface melt pool structure, with improved performance in high energy density regimes compared to analytical melt pool models. The performance of this model is evaluated through dimensional and geometric comparisons to the corresponding experimental melt pool observations.
Authors: Haifeng Yang, Chuanxing Geng, Pong C. Yuen, Songcan Chen
Abstract: In open-set recognition, existing methods generally learn statically fixed decision boundaries using known classes to reject unknown classes. Though they have achieved promising results, such decision boundaries are evidently insufficient for universal unknown classes in dynamic and open scenarios as they can potentially appear at any position in the feature space. Moreover, these methods just simply reject unknown class samples during testing without any effective utilization for them. In fact, such samples completely can constitute the true instantiated representation of the unknown classes to further enhance the model's performance. To address these issues, this paper proposes a novel dynamic against dynamic idea, i.e., dynamic method against dynamic changing open-set world, where an open-set self-learning (OSSL) framework is correspondingly developed. OSSL starts with a good closed-set classifier trained by known classes and utilizes available test samples for model adaptation during testing, thus gaining the adaptability to changing data distributions. In particular, a novel self-matching module is designed for OSSL, which can achieve the adaptation in automatically identifying known class samples while rejecting unknown class samples which are further utilized to enhance the discriminability of the model as the instantiated representation of unknown classes. Our method establishes new performance milestones respectively in almost all standard and cross-data benchmarks.
Authors: Shaily Roy, Harshit Sharma, Asif Salekin
Abstract: Federated learning (FL) enables collaborative model training while preserving data privacy, making it suitable for decentralized human-centered AI applications. However, a significant research gap remains in ensuring fairness in these systems. Current fairness strategies in FL require knowledge of bias-creating/sensitive attributes, clashing with FL's privacy principles. Moreover, in human-centered datasets, sensitive attributes may remain latent. To tackle these challenges, we present a novel bias mitigation approach inspired by "Fairness without Demographics" in machine learning. The presented approach achieves fairness without needing knowledge of sensitive attributes by minimizing the top eigenvalue of the Hessian matrix during training, ensuring equitable loss landscapes across FL participants. Notably, we introduce a novel FL aggregation scheme that promotes participating models based on error rates and loss landscape curvature attributes, fostering fairness across the FL system. This work represents the first approach to attaining "Fairness without Demographics" in human-centered FL. Through comprehensive evaluation, our approach demonstrates effectiveness in balancing fairness and efficacy across various real-world applications, FL setups, and scenarios involving single and multiple bias-inducing factors, representing a significant advancement in human-centered FL.
Authors: Pavamana K J, Chandramani Kishore Singh
Abstract: We study content caching with recommendations in a wireless network where the users are connected through a base station equipped with a finite-capacity cache. We assume a fixed set of contents with unknown user preferences and content popularities. We can recommend a subset of the contents to the users which encourages the users to request these contents. Recommendation can thus be used to increase cache hits. We formulate the cache hit optimization problem as a combinatorial multi-armed bandit (CMAB). We propose a UCB-based algorithm to decide which contents to cache and recommend. We provide an upper bound on the regret of our algorithm. We numerically demonstrate the performance of our algorithm and compare it to state-of-the-art algorithms.
Authors: Youssef Allouah, Anastasia Koloskova, Aymane El Firdoussi, Martin Jaggi, Rachid Guerraoui
Abstract: Decentralized learning is appealing as it enables the scalable usage of large amounts of distributed data and resources (without resorting to any central entity), while promoting privacy since every user minimizes the direct exposure of their data. Yet, without additional precautions, curious users can still leverage models obtained from their peers to violate privacy. In this paper, we propose Decor, a variant of decentralized SGD with differential privacy (DP) guarantees. Essentially, in Decor, users securely exchange randomness seeds in one communication round to generate pairwise-canceling correlated Gaussian noises, which are injected to protect local models at every communication round. We theoretically and empirically show that, for arbitrary connected graphs, Decor matches the central DP optimal privacy-utility trade-off. We do so under SecLDP, our new relaxation of local DP, which protects all user communications against an external eavesdropper and curious users, assuming that every pair of connected users shares a secret, i.e., an information hidden to all others. The main theoretical challenge is to control the accumulation of non-canceling correlated noise due to network sparsity. We also propose a companion SecLDP privacy accountant for public use.
Authors: Mikkel Jordahn, Pablo M. Olmos
Abstract: Deep Neural Networks (DNN) have shown great promise in many classification applications, yet are widely known to have poorly calibrated predictions when they are over-parametrized. Improving DNN calibration without comprising on model accuracy is of extreme importance and interest in safety critical applications such as in the health-care sector. In this work, we show that decoupling the training of feature extraction layers and classification layers in over-parametrized DNN architectures such as Wide Residual Networks (WRN) and Visual Transformers (ViT) significantly improves model calibration whilst retaining accuracy, and at a low training cost. In addition, we show that placing a Gaussian prior on the last hidden layer outputs of a DNN, and training the model variationally in the classification training stage, even further improves calibration. We illustrate these methods improve calibration across ViT and WRN architectures for several image classification benchmark datasets.
Authors: Zhongchang Sun, Sihong He, Fei Miao, Shaofeng Zou
Abstract: Existing studies on constrained reinforcement learning (RL) may obtain a well-performing policy in the training environment. However, when deployed in a real environment, it may easily violate constraints that were originally satisfied during training because there might be model mismatch between the training and real environments. To address the above challenge, we formulate the problem as constrained RL under model uncertainty, where the goal is to learn a good policy that optimizes the reward and at the same time satisfy the constraint under model mismatch. We develop a Robust Constrained Policy Optimization (RCPO) algorithm, which is the first algorithm that applies to large/continuous state space and has theoretical guarantees on worst-case reward improvement and constraint violation at each iteration during the training. We demonstrate the effectiveness of our algorithm on a set of RL tasks with constraints.
Authors: Tianjun Ke, Haoqun Cao, Feng Zhou
Abstract: Bayesian few-shot classification has been a focal point in the field of few-shot learning. This paper seamlessly integrates mirror descent-based variational inference into Gaussian process-based few-shot classification, addressing the challenge of non-conjugate inference. By leveraging non-Euclidean geometry, mirror descent achieves accelerated convergence by providing the steepest descent direction along the corresponding manifold. It also exhibits the parameterization invariance property concerning the variational distribution. Experimental results demonstrate competitive classification accuracy, improved uncertainty quantification, and faster convergence compared to baseline models. Additionally, we investigate the impact of hyperparameters and components. Code is publicly available at https://github.com/keanson/MD-BSFC.
Authors: Luciano Dyballa, Evan Gerritz, Steven W. Zucker
Abstract: Generalization to unseen data remains poorly understood for deep learning classification and foundation models. How can one assess the ability of networks to adapt to new or extended versions of their input space in the spirit of few-shot learning, out-of-distribution generalization, and domain adaptation? Which layers of a network are likely to generalize best? We provide a new method for evaluating the capacity of networks to represent a sampled domain, regardless of whether the network has been trained on all classes in the domain. Our approach is the following: after fine-tuning state-of-the-art pre-trained models for visual classification on a particular domain, we assess their performance on data from related but distinct variations in that domain. Generalization power is quantified as a function of the latent embeddings of unseen data from intermediate layers for both unsupervised and supervised settings. Working throughout all stages of the network, we find that (i) high classification accuracy does not imply high generalizability; and (ii) deeper layers in a model do not always generalize the best, which has implications for pruning. Since the trends observed across datasets are largely consistent, we conclude that our approach reveals (a function of) the intrinsic capacity of the different layers of a model to generalize.
Authors: Matthew Hausknecht, Peter Stone
Abstract: Recent work has shown that deep neural networks are capable of approximating both value functions and policies in reinforcement learning domains featuring continuous state and action spaces. However, to the best of our knowledge no previous work has succeeded at using deep neural networks in structured (parameterized) continuous action spaces. To fill this gap, this paper focuses on learning within the domain of simulated RoboCup soccer, which features a small set of discrete action types, each of which is parameterized with continuous variables. The best learned agent can score goals more reliably than the 2012 RoboCup champion agent. As such, this paper represents a successful extension of deep reinforcement learning to the class of parameterized action space MDPs.
Authors: You Zu, Krishnamurthy Iyer, Haifeng Xu
Abstract: Motivated by information sharing in online platforms, we study repeated persuasion between a sender and a stream of receivers where at each time, the sender observes a payoff-relevant state drawn independently and identically from an unknown distribution, and shares state information with the receivers who each choose an action. The sender seeks to persuade the receivers into taking actions aligned with the sender's preference by selectively sharing state information. However, in contrast to the standard models, neither the sender nor the receivers know the distribution, and the sender has to persuade while learning the distribution on the fly. We study the sender's learning problem of making persuasive action recommendations to achieve low regret against the optimal persuasion mechanism with the knowledge of the distribution. To do this, we first propose and motivate a persuasiveness criterion for the unknown distribution setting that centers robustness as a requirement in the face of uncertainty. Our main result is an algorithm that, with high probability, is robustly-persuasive and achieves $O(\sqrt{T\log T})$ regret, where $T$ is the horizon length. Intuitively, at each time our algorithm maintains a set of candidate distributions, and chooses a signaling mechanism that is simultaneously persuasive for all of them. Core to our proof is a tight analysis about the cost of robust persuasion, which may be of independent interest. We further prove that this regret order is optimal (up to logarithmic terms) by showing that no algorithm can achieve regret better than $\Omega(\sqrt{T})$.
Authors: Gangshan Jing, He Bai, Jemin George, Aranya Chakrabortty, Piyush K. Sharma
Abstract: Recently introduced distributed zeroth-order optimization (ZOO) algorithms have shown their utility in distributed reinforcement learning (RL). Unfortunately, in the gradient estimation process, almost all of them require random samples with the same dimension as the global variable and/or require evaluation of the global cost function, which may induce high estimation variance for large-scale networks. In this paper, we propose a novel distributed zeroth-order algorithm by leveraging the network structure inherent in the optimization objective, which allows each agent to estimate its local gradient by local cost evaluation independently, without use of any consensus protocol. The proposed algorithm exhibits an asynchronous update scheme, and is designed for stochastic non-convex optimization with a possibly non-convex feasible domain based on the block coordinate descent method. The algorithm is later employed as a distributed model-free RL algorithm for distributed linear quadratic regulator design, where a learning graph is designed to describe the required interaction relationship among agents in distributed learning. We provide an empirical validation of the proposed algorithm to benchmark its performance on convergence rate and variance against a centralized ZOO algorithm.
Authors: Siddhant Kharbanda, Atmadeep Banerjee, Devaansh Gupta, Akash Palrecha, Rohit Babbar
Abstract: Automatic annotation of short-text data to a large number of target labels, referred to as Short Text Extreme Classification, has found numerous applications including prediction of related searches and product recommendation. In this paper, we propose a convolutional architecture InceptionXML which is light-weight, yet powerful, and robust to the inherent lack of word-order in short-text queries encountered in search and recommendation. We demonstrate the efficacy of applying convolutions by recasting the operation along the embedding dimension instead of the word dimension as applied in conventional CNNs for text classification. Towards scaling our model to datasets with millions of labels, we also propose SyncXML pipeline which improves upon the shortcomings of the recently proposed dynamic hard-negative mining technique for label short-listing by synchronizing the label-shortlister and extreme classifier. SyncXML not only reduces the inference time to half but is also an order of magnitude smaller than state-of-the-art Astec in terms of model size. Through a comprehensive empirical comparison, we show that not only can InceptionXML outperform existing approaches on benchmark datasets but also the transformer baselines requiring only 2% FLOPs. The code for InceptionXML is available at https://github.com/xmc-aalto/inceptionxml.
Authors: Wasim Huleihel, Arya Mazumdar, Soumyabrata Pal
Abstract: The planted densest subgraph detection problem refers to the task of testing whether in a given (random) graph there is a subgraph that is unusually dense. Specifically, we observe an undirected and unweighted graph on $n$ vertices. Under the null hypothesis, the graph is a realization of an Erd\H{o}s-R\'{e}nyi graph with edge probability (or, density) $q$. Under the alternative, there is a subgraph on $k$ vertices with edge probability $p>q$. The statistical as well as the computational barriers of this problem are well-understood for a wide range of the edge parameters $p$ and $q$. In this paper, we consider a natural variant of the above problem, where one can only observe a relatively small part of the graph using adaptive edge queries. For this model, we determine the number of queries necessary and sufficient (accompanied with a quasi-polynomial optimal algorithm) for detecting the presence of the planted subgraph. We also propose a polynomial-time algorithm which is able to detect the planted subgraph, albeit with more queries compared to the above lower bound. We conjecture that in the leftover regime, no polynomial-time algorithms exist. Our results resolve two open questions posed in the past literature.
Authors: Jon Donnelly, Alina Jade Barnett, Chaofan Chen
Abstract: We present a deformable prototypical part network (Deformable ProtoPNet), an interpretable image classifier that integrates the power of deep learning and the interpretability of case-based reasoning. This model classifies input images by comparing them with prototypes learned during training, yielding explanations in the form of "this looks like that." However, while previous methods use spatially rigid prototypes, we address this shortcoming by proposing spatially flexible prototypes. Each prototype is made up of several prototypical parts that adaptively change their relative spatial positions depending on the input image. Consequently, a Deformable ProtoPNet can explicitly capture pose variations and context, improving both model accuracy and the richness of explanations provided. Compared to other case-based interpretable models using prototypes, our approach achieves state-of-the-art accuracy and gives an explanation with greater context. The code is available at https://github.com/jdonnelly36/Deformable-ProtoPNet.
Authors: Tsung-Wei Ke, Sangwoo Mo, Stella X. Yu
Abstract: Large vision and language models learned directly through image-text associations often lack detailed visual substantiation, whereas image segmentation tasks are treated separately from recognition, supervisedly learned without interconnections. Our key observation is that, while an image can be recognized in multiple ways, each has a consistent part-and-whole visual organization. Segmentation thus should be treated not as an end task to be mastered through supervised learning, but as an internal process that evolves with and supports the ultimate goal of recognition. We propose to integrate a hierarchical segmenter into the recognition process, train and adapt the entire model solely on image-level recognition objectives. We learn hierarchical segmentation for free alongside recognition, automatically uncovering part-to-whole relationships that not only underpin but also enhance recognition. Enhancing the Vision Transformer (ViT) with adaptive segment tokens and graph pooling, our model surpasses ViT in unsupervised part-whole discovery, semantic segmentation, image classification, and efficiency. Notably, our model (trained on unlabeled 1M ImageNet images) outperforms SAM (trained on 11M images and 1 billion masks) by absolute 8% in mIoU on PartImageNet object segmentation.
Authors: Nour Neifar, Achraf Ben-Hamadou, Afef Mdhaffar, Mohamed Jmaiel
Abstract: Within cardiovascular disease detection using deep learning applied to ECG signals, the complexities of handling physiological signals have sparked growing interest in leveraging deep generative models for effective data augmentation. In this paper, we introduce a novel versatile approach based on denoising diffusion probabilistic models for ECG synthesis, addressing three scenarios: (i) heartbeat generation, (ii) partial signal imputation, and (iii) full heartbeat forecasting. Our approach presents the first generalized conditional approach for ECG synthesis, and our experimental results demonstrate its effectiveness for various ECG-related tasks. Moreover, we show that our approach outperforms other state-of-the-art ECG generative models and can enhance the performance of state-of-the-art classifiers.
Authors: Allan Jabri, Sjoerd van Steenkiste, Emiel Hoogeboom, Mehdi S. M. Sajjadi, Thomas Kipf
Abstract: Recent progress in 3D scene understanding enables scalable learning of representations across large datasets of diverse scenes. As a consequence, generalization to unseen scenes and objects, rendering novel views from just a single or a handful of input images, and controllable scene generation that supports editing, is now possible. However, training jointly on a large number of scenes typically compromises rendering quality when compared to single-scene optimized models such as NeRFs. In this paper, we leverage recent progress in diffusion models to equip 3D scene representation learning models with the ability to render high-fidelity novel views, while retaining benefits such as object-level scene editing to a large degree. In particular, we propose DORSal, which adapts a video diffusion architecture for 3D scene generation conditioned on frozen object-centric slot-based representations of scenes. On both complex synthetic multi-object scenes and on the real-world large-scale Street View dataset, we show that DORSal enables scalable neural rendering of 3D scenes with object-level editing and improves upon existing approaches.
Authors: Andrea Angiuli, Jean-Pierre Fouque, Ruimeng Hu, Alan Raydan
Abstract: We present the development and analysis of a reinforcement learning (RL) algorithm designed to solve continuous-space mean field game (MFG) and mean field control (MFC) problems in a unified manner. The proposed approach pairs the actor-critic (AC) paradigm with a representation of the mean field distribution via a parameterized score function, which can be efficiently updated in an online fashion, and uses Langevin dynamics to obtain samples from the resulting distribution. The AC agent and the score function are updated iteratively to converge, either to the MFG equilibrium or the MFC optimum for a given mean field problem, depending on the choice of learning rates. A straightforward modification of the algorithm allows us to solve mixed mean field control games (MFCGs). The performance of our algorithm is evaluated using linear-quadratic benchmarks in the asymptotic infinite horizon framework.
Authors: Lukas Schneider, Jonas Frey, Takahiro Miki, Marco Hutter
Abstract: Deployment in hazardous environments requires robots to understand the risks associated with their actions and movements to prevent accidents. Despite its importance, these risks are not explicitly modeled by currently deployed locomotion controllers for legged robots. In this work, we propose a risk sensitive locomotion training method employing distributional reinforcement learning to consider safety explicitly. Instead of relying on a value expectation, we estimate the complete value distribution to account for uncertainty in the robot's interaction with the environment. The value distribution is consumed by a risk metric to extract risk sensitive value estimates. These are integrated into Proximal Policy Optimization (PPO) to derive our method, Distributional Proximal Policy Optimization (DPPO). The risk preference, ranging from risk-averse to risk-seeking, can be controlled by a single parameter, which enables to adjust the robot's behavior dynamically. Importantly, our approach removes the need for additional reward function tuning to achieve risk sensitivity. We show emergent risk sensitive locomotion behavior in simulation and on the quadrupedal robot ANYmal. Videos of the experiments and code are available at https://sites.google.com/leggedrobotics.com/risk-aware-locomotion.
URLs: https://sites.google.com/leggedrobotics.com/risk-aware-locomotion.
Authors: Olaf Lipinski, Adam J. Sobey, Federico Cerutti, Timothy J. Norman
Abstract: Emergent communication studies the development of language between autonomous agents, aiming to improve understanding of natural language evolution and increase communication efficiency. While temporal aspects of language have been considered in computational linguistics, there has been no research on temporal references in emergent communication. This paper addresses this gap, by exploring how agents communicate about temporal relationships. We analyse three potential influences for the emergence of temporal references: environmental, external, and architectural changes. Our experiments demonstrate that altering the loss function is insufficient for temporal references to emerge; rather, architectural changes are necessary. However, a minimal change in agent architecture, using a different batching method, allows the emergence of temporal references. This modified design is compared with the standard architecture in a temporal referential games environment, which emphasises temporal relationships. The analysis indicates that over 95\% of the agents with the modified batching method develop temporal references, without changes to their loss function. We consider temporal referencing necessary for future improvements to the agents' communication efficiency, yielding a closer to optimal coding as compared to purely compositional languages. Our readily transferable architectural insights provide the basis for their incorporation into other emergent communication settings.
Authors: Jae Hee Lee, Sergio Lanza, Stefan Wermter
Abstract: In this paper, we review recent approaches for explaining concepts in neural networks. Concepts can act as a natural link between learning and reasoning: once the concepts are identified that a neural learning system uses, one can integrate those concepts with a reasoning system for inference or use a reasoning system to act upon them to improve or enhance the learning system. On the other hand, knowledge can not only be extracted from neural networks but concept knowledge can also be inserted into neural network architectures. Since integrating learning and reasoning is at the core of neuro-symbolic AI, the insights gained from this survey can serve as an important step towards realizing neuro-symbolic AI based on explainable concepts.
Authors: R\=uta Binkyt\.e, Carlos Pinz\'on, Szilvia Lesty\'an, Kangsoo Jung, H\'eber H. Arcolezi, Catuscia Palamidessi
Abstract: Differential privacy is a widely adopted framework designed to safeguard the sensitive information of data providers within a data set. It is based on the application of controlled noise at the interface between the server that stores and processes the data, and the data consumers. Local differential privacy is a variant that allows data providers to apply the privatization mechanism themselves on their data individually. Therefore it provides protection also in contexts in which the server, or even the data collector, cannot be trusted. The introduction of noise, however, inevitably affects the utility of the data, particularly by distorting the correlations between individual data components. This distortion can prove detrimental to tasks such as causal discovery. In this paper, we consider various well-known locally differentially private mechanisms and compare the trade-off between the privacy they provide, and the accuracy of the causal structure produced by algorithms for causal learning when applied to data obfuscated by these mechanisms. Our analysis yields valuable insights for selecting appropriate local differentially private protocols for causal discovery tasks. We foresee that our findings will aid researchers and practitioners in conducting locally private causal discovery.
Authors: Julian Berberich, Daniel Fink, Daniel Pranji\'c, Christian Tutschku, Christian Holm
Abstract: Adversarial robustness and generalization are both crucial properties of reliable machine learning models. In this letter, we study these properties in the context of quantum machine learning based on Lipschitz bounds. We derive parameter-dependent Lipschitz bounds for quantum models with trainable encoding, showing that the norm of the data encoding has a crucial impact on the robustness against data perturbations. Further, we derive a bound on the generalization error which explicitly involves the parameters of the data encoding. Our theoretical findings give rise to a practical strategy for training robust and generalizable quantum models by regularizing the Lipschitz bound in the cost. Further, we show that, for fixed and non-trainable encodings, as those frequently employed in quantum machine learning, the Lipschitz bound cannot be influenced by tuning the parameters. Thus, trainable encodings are crucial for systematically adapting robustness and generalization during training. The practical implications of our theoretical findings are illustrated with numerical results.
Authors: Micha Livne, Zulfat Miftahutdinov, Elena Tutubalina, Maksim Kuznetsov, Daniil Polykovskiy, Annika Brundyn, Aastha Jhunjhunwala, Anthony Costa, Alex Aliper, Al\'an Aspuru-Guzik, Alex Zhavoronkov
Abstract: Large Language Models (LLMs) have substantially driven scientific progress in various domains, and many papers have demonstrated their ability to tackle complex problems with creative solutions. Our paper introduces a new foundation model, nach0, capable of solving various chemical and biological tasks: biomedical question answering, named entity recognition, molecular generation, molecular synthesis, attributes prediction, and others. nach0 is a multi-domain and multi-task encoder-decoder LLM pre-trained on unlabeled text from scientific literature, patents, and molecule strings to incorporate a range of chemical and linguistic knowledge. We employed instruction tuning, where specific task-related instructions are utilized to fine-tune nach0 for the final set of tasks. To train nach0 effectively, we leverage the NeMo framework, enabling efficient parallel optimization of both base and large model versions. Extensive experiments demonstrate that our model outperforms state-of-the-art baselines on single-domain and cross-domain tasks. Furthermore, it can generate high-quality outputs in molecular and textual formats, showcasing its effectiveness in multi-domain setups.
Authors: Lijie Hu, Yixin Liu, Ninghao Liu, Mengdi Huai, Lichao Sun, Di Wang
Abstract: Vision Transformers (ViTs) have achieved state-of-the-art performance for various vision tasks. One reason behind the success lies in their ability to provide plausible innate explanations for the behavior of neural architectures. However, ViTs suffer from issues with explanation faithfulness, as their focal points are fragile to adversarial attacks and can be easily changed with even slight perturbations on the input image. In this paper, we propose a rigorous approach to mitigate these issues by introducing Faithful ViTs (FViTs). Briefly speaking, an FViT should have the following two properties: (1) The top-$k$ indices of its self-attention vector should remain mostly unchanged under input perturbation, indicating stable explanations; (2) The prediction distribution should be robust to perturbations. To achieve this, we propose a new method called Denoised Diffusion Smoothing (DDS), which adopts randomized smoothing and diffusion-based denoising. We theoretically prove that processing ViTs directly with DDS can turn them into FViTs. We also show that Gaussian noise is nearly optimal for both $\ell_2$ and $\ell_\infty$-norm cases. Finally, we demonstrate the effectiveness of our approach through comprehensive experiments and evaluations. Results show that FViTs are more robust against adversarial attacks while maintaining the explainability of attention, indicating higher faithfulness.
Authors: James Seale Smith, Yen-Chang Hsu, Zsolt Kira, Yilin Shen, Hongxia Jin
Abstract: Recent work has demonstrated a remarkable ability to customize text-to-image diffusion models to multiple, fine-grained concepts in a sequential (i.e., continual) manner while only providing a few example images for each concept. This setting is known as continual diffusion. Here, we ask the question: Can we scale these methods to longer concept sequences without forgetting? Although prior work mitigates the forgetting of previously learned concepts, we show that its capacity to learn new tasks reaches saturation over longer sequences. We address this challenge by introducing a novel method, STack-And-Mask INcremental Adapters (STAMINA), which is composed of low-ranked attention-masked adapters and customized MLP tokens. STAMINA is designed to enhance the robust fine-tuning properties of LoRA for sequential concept learning via learnable hard-attention masks parameterized with low rank MLPs, enabling precise, scalable learning via sparse adaptation. Notably, all introduced trainable parameters can be folded back into the model after training, inducing no additional inference parameter costs. We show that STAMINA outperforms the prior SOTA for the setting of text-to-image continual customization on a 50-concept benchmark composed of landmarks and human faces, with no stored replay data. Additionally, we extended our method to the setting of continual learning for image classification, demonstrating that our gains also translate to state-of-the-art performance in this standard benchmark.
Authors: Matteo Gioele Collu, Tom Janssen-Groesbeek, Stefanos Koffas, Mauro Conti, Stjepan Picek
Abstract: Recently, we have witnessed a rise in the use of Large Language Models (LLMs), especially in applications like chatbot assistants. Safety mechanisms and specialized training procedures are implemented to prevent improper responses from these assistants. In this work, we bypass these measures for ChatGPT and Bard (and, to some extent, Bing chat) by making them impersonate complex personas with personality characteristics that are not aligned with a truthful assistant. We start by creating elaborate biographies of these personas, which we then use in a new session with the same chatbots. Our conversations then followed a role-play style to elicit prohibited responses. By making use of personas, we show that such responses are actually provided, making it possible to obtain unauthorized, illegal, or harmful information. This work shows that by using adversarial personas, one can overcome safety mechanisms set out by ChatGPT and Bard. We also introduce several ways of activating such adversarial personas, which show that both chatbots are vulnerable to this kind of attack. With the same principle, we introduce two defenses that push the model to interpret trustworthy personalities and make it more robust against such attacks.
Authors: Giulia Baldini, Melanie Schmidt, Charlotte Z\"aske, Liliana L. Caldeira
Abstract: We consider a missing data problem in the context of automatic segmentation methods for Magnetic Resonance Imaging (MRI) brain scans. Usually, automated MRI scan segmentation is based on multiple scans (e.g., T1-weighted, T2-weighted, T1CE, FLAIR). However, quite often a scan is blurry, missing or otherwise unusable. We investigate the question whether a missing scan can be synthesized. We exemplify that this is in principle possible by synthesizing a T2-weighted scan from a given T1-weighted scan. Our first aim is to compute a picture that resembles the missing scan closely, measured by average mean squared error (MSE). We develop/use several methods for this, including a random baseline approach, a clustering-based method and pixel-to-pixel translation method by Isola et al. (Pix2Pix) which is based on conditional GANs. The lowest MSE is achieved by our clustering-based method. Our second aim is to compare the methods with respect to the effect that using the synthesized scan has on the segmentation process. For this, we use a DeepMedic model trained with the four input scan modalities named above. We replace the T2-weighted scan by the synthesized picture and evaluate the segmentations with respect to the tumor identification, using Dice scores as numerical evaluation. The evaluation shows that the segmentation works well with synthesized scans (in particular, with Pix2Pix methods) in many cases.
Authors: Shu Ishida, Gianluca Corrado, George Fedoseev, Hudson Yeo, Lloyd Russell, Jamie Shotton, Jo\~ao F. Henriques, Anthony Hu
Abstract: We propose LangProp, a framework for iteratively optimizing code generated by large language models (LLMs), in both supervised and reinforcement learning settings. While LLMs can generate sensible coding solutions zero-shot, they are often sub-optimal. Especially for code generation tasks, it is likely that the initial code will fail on certain edge cases. LangProp automatically evaluates the code performance on a dataset of input-output pairs, catches any exceptions, and feeds the results back to the LLM in the training loop, so that the LLM can iteratively improve the code it generates. By adopting a metric- and data-driven training paradigm for this code optimization procedure, one could easily adapt findings from traditional machine learning techniques such as imitation learning, DAgger, and reinforcement learning. We show LangProp's applicability to general domains such as Sudoku and CartPole, as well as demonstrate the first proof of concept of automated code optimization for autonomous driving in CARLA. We show that LangProp can generate interpretable and transparent policies that can be verified and improved in a metric- and data-driven way. Our code is available at https://github.com/shuishida/LangProp.
Authors: Carlos Oliver, Vincent Mallet, J\'er\^ome Waldisp\"uhl
Abstract: Understanding the connection between complex structural features of RNA and biological function is a fundamental challenge in evolutionary studies and in RNA design. However, building datasets of RNA 3D structures and making appropriate modeling choices remains time-consuming and lacks standardization. In this chapter, we describe the use of rnaglib, to train supervised and unsupervised machine learning-based function prediction models on datasets of RNA 3D structures.
Authors: Evan Gerritz, Luciano Dyballa, Steven W. Zucker
Abstract: Generalization to unseen data is a key desideratum for deep networks, but its relation to classification accuracy is unclear. Using a minimalist vision dataset and a measure of generalizability, we show that popular networks, from deep convolutional networks (CNNs) to transformers, vary in their power to extrapolate to unseen classes both across layers and across architectures. Accuracy is not a good predictor of generalizability, and generalization varies non-monotonically with layer depth.
Authors: Philipp Schoenegger, Indre Tuminauskaite, Peter S. Park, Philip E. Tetlock
Abstract: Human forecasting accuracy in practice relies on the 'wisdom of the crowd' effect, in which predictions about future events are significantly improved by aggregating across a crowd of individual forecasters. Past work on the forecasting ability of large language models (LLMs) suggests that frontier LLMs, as individual forecasters, underperform compared to the gold standard of a human crowd forecasting tournament aggregate. In Study 1, we expand this research by using an LLM ensemble approach consisting of a crowd of twelve LLMs. We compare the aggregated LLM predictions on 31 binary questions to that of a crowd of 925 human forecasters from a three-month forecasting tournament. Our preregistered main analysis shows that the LLM crowd outperforms a simple no-information benchmark and is not statistically different from the human crowd. In exploratory analyses, we find that these two approaches are equivalent with respect to medium-effect-size equivalence bounds. We also observe an acquiescence effect, with mean model predictions being significantly above 50%, despite an almost even split of positive and negative resolutions. Moreover, in Study 2, we test whether LLM predictions (of GPT-4 and Claude 2) can be improved by drawing on human cognitive output. We find that both models' forecasting accuracy benefits from exposure to the median human prediction as information, improving accuracy by between 17% and 28%: though this leads to less accurate predictions than simply averaging human and machine forecasts. Our results suggest that LLMs can achieve forecasting accuracy rivaling that of human crowd forecasting tournaments: via the simple, practically applicable method of forecast aggregation. This replicates the 'wisdom of the crowd' effect for LLMs, and opens up their use for a variety of applications throughout society.
Authors: Philip W. Livermore, Leyuan Wu, Longwei Chen, Sjoerd A. L. de Ridder
Abstract: Magnetic sounding using data collected from the Juno mission can be used to provide constraints on Jupiter's interior. However, inwards continuation of reconstructions assuming zero electrical conductivity and a representation in spherical harmonics are limited by the enhancement of noise at small scales. Here we describe new reconstructions of Jupiter's internal magnetic field based on physics-informed neural networks and either the first 33 (PINN33) or the first 50 (PINN50) of Juno's orbits. The method can resolve local structures, and allows for weak ambient electrical currents. Our models are not hampered by noise amplification at depth, and offer a much clearer picture of the interior structure. We estimate that the dynamo boundary is at a fractional radius of 0.8. At this depth, the magnetic field is arranged into longitudinal bands, and strong local features such as the great blue spot appear to be rooted in neighbouring structures of oppositely signed flux.
Authors: Thennal D K, Ganesh Nathan, Suchithra M S
Abstract: Fine-tuning pre-trained models provides significant advantages in downstream performance. The ubiquitous nature of pre-trained models such as BERT and its derivatives in natural language processing has also led to a proliferation of task-specific fine-tuned models. As these models typically only perform one task well, additional training or ensembling is required in multi-task scenarios. The growing field of model merging provides a solution, dealing with the challenge of combining multiple task-specific models into a single multi-task model. In this study, we introduce a novel model merging method for Transformers, combining insights from previous work in Fisher-weighted averaging and the use of Fisher information in model pruning. Utilizing the Fisher information of mask nodes within the Transformer architecture, we devise a computationally efficient weighted-averaging scheme. Our method exhibits a regular and significant performance increase across various models in the BERT family, outperforming full-scale Fisher-weighted averaging in a fraction of the computational cost, with baseline performance improvements of up to +6.5 and a speedup between 57.4x and 321.7x across models. Our results prove the potential of our method in current multi-task learning environments and suggest its scalability and adaptability to new model architectures and learning scenarios.
Authors: Guangming Huang, Yingya Li, Shoaib Jameel, Yunfei Long, Giorgos Papanastasiou
Abstract: Deep learning (DL) has substantially enhanced natural language processing (NLP) in healthcare research. However, the increasing complexity of DL-based NLP necessitates transparent model interpretability, or at least explainability, for reliable decision-making. This work presents a thorough scoping review of explainable and interpretable DL in healthcare NLP. The term "eXplainable and Interpretable Artificial Intelligence" (XIAI) is introduced to distinguish XAI from IAI. Different models are further categorized based on their functionality (model-, input-, output-based) and scope (local, global). Our analysis shows that attention mechanisms are the most prevalent emerging IAI technique. The use of IAI is growing, distinguishing it from XAI. The major challenges identified are that most XIAI does not explore "global" modelling processes, the lack of best practices, and the lack of systematic evaluation and benchmarks. One important opportunity is to use attention mechanisms to enhance multi-modal XIAI for personalized medicine. Additionally, combining DL with causal logic holds promise. Our discussion encourages the integration of XIAI in Large Language Models (LLMs) and domain-specific smaller models. In conclusion, XIAI adoption in healthcare requires dedicated in-house expertise. Collaboration with domain experts, end-users, and policymakers can lead to ready-to-use XIAI methods across NLP and medical tasks. While challenges exist, XIAI techniques offer a valuable foundation for interpretable NLP algorithms in healthcare.
Authors: Hao Tang, Lianglun Cheng, Guoheng Huang, Zhengguang Tan, Junhao Lu, Kaihong Wu
Abstract: Image segmentation holds a vital position in the realms of diagnosis and treatment within the medical domain. Traditional convolutional neural networks (CNNs) and Transformer models have made significant advancements in this realm, but they still encounter challenges because of limited receptive field or high computing complexity. Recently, State Space Models (SSMs), particularly Mamba and its variants, have demonstrated notable performance in the field of vision. However, their feature extraction methods may not be sufficiently effective and retain some redundant structures, leaving room for parameter reduction. Motivated by previous spatial and channel attention methods, we propose Triplet Mamba-UNet. The method leverages residual VSS Blocks to extract intensive contextual features, while Triplet SSM is employed to fuse features across spatial and channel dimensions. We conducted experiments on ISIC17, ISIC18, CVC-300, CVC-ClinicDB, Kvasir-SEG, CVC-ColonDB, and Kvasir-Instrument datasets, demonstrating the superior segmentation performance of our proposed TM-UNet. Additionally, compared to the previous VM-UNet, our model achieves a one-third reduction in parameters.
Authors: Manav Mishra, Hritik Bana, Saswata Sarkar, Sujeevraja Sanjeevi, PB Sujit, Kaarthik Sundar
Abstract: This article presents a deep reinforcement learning-based approach to tackle a persistent surveillance mission requiring a single unmanned aerial vehicle initially stationed at a depot with fuel or time-of-flight constraints to repeatedly visit a set of targets with equal priority. Owing to the vehicle's fuel or time-of-flight constraints, the vehicle must be regularly refueled, or its battery must be recharged at the depot. The objective of the problem is to determine an optimal sequence of visits to the targets that minimizes the maximum time elapsed between successive visits to any target while ensuring that the vehicle never runs out of fuel or charge. We present a deep reinforcement learning algorithm to solve this problem and present the results of numerical experiments that corroborate the effectiveness of this approach in comparison with common-sense greedy heuristics.
Authors: Cailean Osborne
Abstract: Governments are increasingly funding open source software (OSS) development to support software security, digital sovereignty, and national competitiveness in science and innovation, amongst others. However, little is known about how OSS developers evaluate the relative benefits and drawbacks of governmental funding for OSS. This study explores this question through a case study on scikit-learn, a Python library for machine learning, funded by public research grants, commercial sponsorship, micro-donations, and a 32 euro million grant announced in France's artificial intelligence strategy. Through 25 interviews with scikit-learn's maintainers and funders, this study makes two key contributions. First, it contributes empirical findings about the benefits and drawbacks of public and private funding in an impactful OSS project, and the governance protocols employed by the maintainers to balance the diverse interests of their community and funders. Second, it offers practical lessons on funding for OSS developers, governments, and companies based on the experience of scikit-learn. The paper concludes with key recommendations for practitioners and future research directions.
Authors: Anton Sch\"afer, Thomas Hofmann, Imanol Schlag, Tiago Pimentel
Abstract: Tokenisation is a core part of language models (LMs). It involves splitting a character sequence into subwords which are assigned arbitrary indices before being served to the LM. While typically lossless, however, this process may lead to less sample efficient LM training: as it removes character-level information, it could make it harder for LMs to generalise across similar subwords, such as now and Now. We refer to such subwords as near duplicates. In this paper, we study the impact of near duplicate subwords on LM training efficiency. First, we design an experiment that gives us an upper bound to how much we should expect a model to improve if we could perfectly generalise across near duplicates. We do this by duplicating each subword in our LM's vocabulary, creating perfectly equivalent classes of subwords. Experimentally, we find that LMs need roughly 17% more data when trained in a fully duplicated setting. Second, we investigate the impact of naturally occurring near duplicates on LMs. Here, we see that merging them considerably hurts LM performance. Therefore, although subword duplication negatively impacts LM training efficiency, naturally occurring near duplicates may not be as similar as anticipated, limiting the potential for performance improvements.
Authors: Keqin Li, Armando Zhu, Peng Zhao, Jintong Song, Jiabei Liu
Abstract: This study explores the application of deep learning technologies in software development processes, particularly in automating code reviews, error prediction, and test generation to enhance code quality and development efficiency. Through a series of empirical studies, experimental groups using deep learning tools and control groups using traditional methods were compared in terms of code error rates and project completion times. The results demonstrated significant improvements in the experimental group, validating the effectiveness of deep learning technologies. The research also discusses potential optimization points, methodologies, and technical challenges of deep learning in software development, as well as how to integrate these technologies into existing software development workflows.
Authors: Boyang Sun, Yu Yao, Huangyuan Hao, Yumou Qiu, Kun Zhang
Abstract: Testing conditional independence has many applications, such as in Bayesian network learning and causal discovery. Different test methods have been proposed. However, existing methods generally can not work when only discretized observations are available. Specifically, consider $X_1$, $\tilde{X}_2$ and $X_3$ are observed variables, where $\tilde{X}_2$ is a discretization of latent variables $X_2$. Applying existing test methods to the observations of $X_1$, $\tilde{X}_2$ and $X_3$ can lead to a false conclusion about the underlying conditional independence of variables $X_1$, $X_2$ and $X_3$. Motivated by this, we propose a conditional independence test specifically designed to accommodate the presence of such discretization. To achieve this, we design the bridge equations to recover the parameter reflecting the statistical information of the underlying latent continuous variables. An appropriate test statistic and its asymptotic distribution under the null hypothesis of conditional independence have also been derived. Both theoretical results and empirical validation have been provided, demonstrating the effectiveness of our test methods.
Authors: Ian W. McBrearty, Paul Segall
Abstract: During the three month long eruption of Kilauea volcano, Hawaii in 2018, the pre-existing summit caldera collapsed in over 60 quasi-periodic failure events. The last 40 of these events, which generated Mw >5 very long period (VLP) earthquakes, had inter-event times between 0.8 - 2.2 days. These failure events offer a unique dataset for testing methods for predicting earthquake recurrence based on locally recorded GPS, tilt, and seismicity data. In this work, we train a deep learning graph neural network (GNN) to predict the time-to-failure of the caldera collapse events using only a fraction of the data recorded at the start of each cycle. We find that the GNN generalizes to unseen data and can predict the time-to-failure to within a few hours using only 0.5 days of data, substantially improving upon a null model based only on inter-event statistics. Predictions improve with increasing input data length, and are most accurate when using high-SNR tilt-meter data. Applying the trained GNN to synthetic data with different magma pressure decay times predicts failure at a nearly constant stress threshold, revealing that the GNN is sensing the underling physics of caldera collapse. These findings demonstrate the predictability of caldera collapse sequences under well monitored conditions, and highlight the potential of machine learning methods for forecasting real world catastrophic events with limited training data.
Authors: Hugh Zhang, Jeff Da, Dean Lee, Vaughn Robinson, Catherine Wu, Will Song, Tiffany Zhao, Pranav Raja, Dylan Slack, Qin Lyu, Sean Hendryx, Russell Kaplan, Michele Lunati, Summer Yue
Abstract: Large language models (LLMs) have achieved impressive success on many benchmarks for mathematical reasoning. However, there is growing concern that some of this performance actually reflects dataset contamination, where data closely resembling benchmark questions leaks into the training data, instead of true reasoning ability. To investigate this claim rigorously, we commission Grade School Math 1000 (GSM1k). GSM1k is designed to mirror the style and complexity of the established GSM8k benchmark, the gold standard for measuring elementary mathematical reasoning. We ensure that the two benchmarks are comparable across important metrics such as human solve rates, number of steps in solution, answer magnitude, and more. When evaluating leading open- and closed-source LLMs on GSM1k, we observe accuracy drops of up to 13%, with several families of models (e.g., Phi and Mistral) showing evidence of systematic overfitting across almost all model sizes. At the same time, many models, especially those on the frontier, (e.g., Gemini/GPT/Claude) show minimal signs of overfitting. Further analysis suggests a positive relationship (Spearman's r^2=0.32) between a model's probability of generating an example from GSM8k and its performance gap between GSM8k and GSM1k, suggesting that many models may have partially memorized GSM8k.