Authors: Kamyar Zeinalipour, Marco Gori
The electrocardiogram (ECG) is a dependable instrument for assessing the function of the cardiovascular system. There has recently been much emphasis on precisely classifying ECGs. While ECG situations have numerous similarities, little attention has been paid to categorizing ECGs using graph neural networks. In this study, we offer three distinct techniques for classifying heartbeats using deep graph neural networks to classify the ECG signals accurately. We suggest using different methods to extract topological features from the ECG signal and then using a branch of the graph neural network named graph isomorphism network for classifying the ECGs. On the PTB Diagnostics data set, we tested the three proposed techniques. According to the findings, the three proposed techniques are capable of making arrhythmia classification predictions with the accuracy of 99.38, 98.76, and 91.93 percent, respectively.
Authors: Norman Mu, Sarah Chen, Zifan Wang, Sizhe Chen, David Karamardian, Lulwa Aljeraisy, Dan Hendrycks, David Wagner
As Large Language Models (LLMs) are deployed with increasing real-world responsibilities, it is important to be able to specify and constrain the behavior of these systems in a reliable manner. Model developers may wish to set explicit rules for the model, such as "do not generate abusive content", but these may be circumvented by jailbreaking techniques. Evaluating how well LLMs follow developer-provided rules in the face of adversarial inputs typically requires manual review, which slows down monitoring and methods development. To address this issue, we propose Rule-following Language Evaluation Scenarios (RuLES), a programmatic framework for measuring rule-following ability in LLMs. RuLES consists of 15 simple text scenarios in which the model is instructed to obey a set of rules in natural language while interacting with the human user. Each scenario has a concise evaluation program to determine whether the model has broken any rules in a conversation. Through manual exploration of model behavior in our scenarios, we identify 6 categories of attack strategies and collect two suites of test cases: one consisting of unique conversations from manual testing and one that systematically implements strategies from the 6 categories. Across various popular proprietary and open models such as GPT-4 and Llama 2, we find that all models are susceptible to a wide variety of adversarial hand-crafted user inputs, though GPT-4 is the best-performing model. Additionally, we evaluate open models under gradient-based attacks and find significant vulnerabilities. We propose RuLES as a challenging new setting for research into exploring and defending against both manual and automatic attacks on LLMs.
Authors: Sang-Hyun Je, Wontae Choi, Kwangjin Oh
The goal of knowledge graph completion (KGC) is to predict missing links in a KG using trained facts that are already known. In recent, pre-trained language model (PLM) based methods that utilize both textual and structural information are emerging, but their performances lag behind state-of-the-art (SOTA) structure-based methods or some methods lose their inductive inference capabilities in the process of fusing structure embedding to text encoder. In this paper, we propose a novel method to effectively unify structure information and language semantics without losing the power of inductive reasoning. We adopt entity anchors and these anchors and textual description of KG elements are fed together into the PLM-based encoder to learn unified representations. In addition, the proposed method utilizes additional random negative samples which can be reused in the each mini-batch during contrastive learning to learn a generalized entity representations. We verify the effectiveness of the our proposed method through various experiments and analysis. The experimental results on standard benchmark widely used in link prediction task show that the proposed model outperforms existing the SOTA KGC models. Especially, our method show the largest performance improvement on FB15K-237, which is competitive to the SOTA of structure-based KGC methods.
Authors: Qinghao Ye, Haiyang Xu, Jiabo Ye, Ming Yan, Haowei Liu, Qi Qian, Ji Zhang, Fei Huang, Jingren Zhou
Multi-modal Large Language Models (MLLMs) have demonstrated impressive instruction abilities across various open-ended tasks. However, previous methods primarily focus on enhancing multi-modal capabilities. In this work, we introduce a versatile multi-modal large language model, mPLUG-Owl2, which effectively leverages modality collaboration to improve performance in both text and multi-modal tasks. mPLUG-Owl2 utilizes a modularized network design, with the language decoder acting as a universal interface for managing different modalities. Specifically, mPLUG-Owl2 incorporates shared functional modules to facilitate modality collaboration and introduces a modality-adaptive module that preserves modality-specific features. Extensive experiments reveal that mPLUG-Owl2 is capable of generalizing both text tasks and multi-modal tasks and achieving state-of-the-art performances with a single generic model. Notably, mPLUG-Owl2 is the first MLLM model that demonstrates the modality collaboration phenomenon in both pure-text and multi-modal scenarios, setting a pioneering path in the development of future multi-modal foundation models.
Authors: Motonari Kambara, Komei Sugiura
This paper aims to develop a framework that enables a robot to execute tasks based on visual information, in response to natural language instructions for Fetch-and-Carry with Object Grounding (FCOG) tasks. Although there have been many frameworks, they usually rely on manually given instruction sentences. Therefore, evaluations have only been conducted with fixed tasks. Furthermore, many multimodal language understanding models for the benchmarks only consider discrete actions. To address the limitations, we propose a framework for the full automation of the generation, execution, and evaluation of FCOG tasks. In addition, we introduce an approach to solving the FCOG tasks by dividing them into four distinct subtasks.
Authors: Angelika Romanou, Syrielle Montariol, Debjit Paul, Leo Laugier, Karl Aberer, Antoine Bosselut
Understanding narratives requires reasoning about the cause-and-effect relationships between events mentioned in the text. While existing foundation models yield impressive results in many NLP tasks requiring reasoning, it is unclear whether they understand the complexity of the underlying network of causal relationships of events in narratives. In this work, we present CRAB, a new Causal Reasoning Assessment Benchmark designed to evaluate causal understanding of events in real-world narratives. CRAB contains fine-grained, contextual causality annotations for ~2.7K pairs of real-world events that describe various newsworthy event timelines (e.g., the acquisition of Twitter by Elon Musk). Using CRAB, we measure the performance of several large language models, demonstrating that most systems achieve poor performance on the task. Motivated by classical causal principles, we also analyze the causal structures of groups of events in CRAB, and find that models perform worse on causal reasoning when events are derived from complex causal structures compared to simple linear causal chains. We make our dataset and code available to the research community.
Authors: Zhongfen Deng, Seunghyun Yoon, Trung Bui, Franck Dernoncourt, Quan Hung Tran, Shuaiqi Liu, Wenting Zhao, Tao Zhang, Yibo Wang, Philip S. Yu
Aspect-based meeting transcript summarization aims to produce multiple summaries, each focusing on one aspect of content in a meeting transcript. It is challenging as sentences related to different aspects can mingle together, and those relevant to a specific aspect can be scattered throughout the long transcript of a meeting. The traditional summarization methods produce one summary mixing information of all aspects, which cannot deal with the above challenges of aspect-based meeting transcript summarization. In this paper, we propose a two-stage method for aspect-based meeting transcript summarization. To select the input content related to specific aspects, we train a sentence classifier on a dataset constructed from the AMI corpus with pseudo-labeling. Then we merge the sentences selected for a specific aspect as the input for the summarizer to produce the aspect-based summary. Experimental results on the AMI corpus outperform many strong baselines, which verifies the effectiveness of our proposed method.
Authors: Ryan Cotterell, Anej Svete, Clara Meister, Tianyu Liu, Li Du
Large language models have become one of the most commonly deployed NLP inventions. In the past half-decade, their integration into core natural language processing tools has dramatically increased the performance of such tools, and they have entered the public discourse surrounding artificial intelligence. Consequently, it is important for both developers and researchers alike to understand the mathematical foundations of large language models, as well as how to implement them. These notes are the accompaniment to the theoretical portion of the ETH Z\"urich course on large language models, covering what constitutes a language model from a formal, theoretical perspective.
Authors: Sihao Chen, Hongming Zhang, Tong Chen, Ben Zhou, Wenhao Yu, Dian Yu, Baolin Peng, Hongwei Wang, Dan Roth, Dong Yu
We introduce sub-sentence encoder, a contrastively-learned contextual embedding model for fine-grained semantic representation of text. In contrast to the standard practice with sentence embeddings, where the meaning of an entire sequence of text is encoded into a fixed-length vector, the sub-sentence encoder learns to produce distinct contextual embeddings corresponding to different atomic propositions, i.e. atomic units of meaning expressed within a text sequence. The sub-sentence embeddings are contrastively learned to recognize (inferred) semantic equivalence between propositions across different text sequences. Our experiments show the effectiveness of sub-sentence encoders in applications, such as retrieving supporting facts for fine-grained text attribution or recognizing the conditional semantic similarity between texts. In practice, we demonstrate that sub-sentence encoders keep the same level of inference cost and space complexity compared to sentence encoders.
Authors: Wenbo Zhang, Hangzhi Guo, Ian D Kivlichan, Vinodkumar Prabhakaran, Davis Yadav, Amulya Yadav
Toxicity is an increasingly common and severe issue in online spaces. Consequently, a rich line of machine learning research over the past decade has focused on computationally detecting and mitigating online toxicity. These efforts crucially rely on human-annotated datasets that identify toxic content of various kinds in social media texts. However, such annotations historically yield low inter-rater agreement, which was often dealt with by taking the majority vote or other such approaches to arrive at a single ground truth label. Recent research has pointed out the importance of accounting for the subjective nature of this task when building and utilizing these datasets, and this has triggered work on analyzing and better understanding rater disagreements, and how they could be effectively incorporated into the machine learning developmental pipeline. While these efforts are filling an important gap, there is a lack of a broader framework about the root causes of rater disagreement, and therefore, we situate this work within that broader landscape. In this survey paper, we analyze a broad set of literature on the reasons behind rater disagreements focusing on online toxicity, and propose a detailed taxonomy for the same. Further, we summarize and discuss the potential solutions targeting each reason for disagreement. We also discuss several open issues, which could promote the future development of online toxicity research.
Authors: Sai Munikoti, Anurag Acharya, Sridevi Wagle, Sameera Horawalavithana
Despite the dramatic progress in Large Language Model (LLM) development, LLMs often provide seemingly plausible but not factual information, often referred to as hallucinations. Retrieval-augmented LLMs provide a non-parametric approach to solve these issues by retrieving relevant information from external data sources and augment the training process. These models help to trace evidence from an externally provided knowledge base allowing the model predictions to be better interpreted and verified. In this work, we critically evaluate these models in their ability to perform in scientific document reasoning tasks. To this end, we tuned multiple such model variants with science-focused instructions and evaluated them on a scientific document reasoning benchmark for the usefulness of the retrieved document passages. Our findings suggest that models justify predictions in science tasks with fabricated evidence and leveraging scientific corpus as pretraining data does not alleviate the risk of evidence fabrication.
Authors: Michael A. Lepori, Thomas Serre, Ellie Pavlick
Neural network models have achieved high performance on a wide variety of complex tasks, but the algorithms that they implement are notoriously difficult to interpret. In order to understand these algorithms, it is often necessary to hypothesize intermediate variables involved in the network's computation. For example, does a language model depend on particular syntactic properties when generating a sentence? However, existing analysis tools make it difficult to test hypotheses of this type. We propose a new analysis technique -- circuit probing -- that automatically uncovers low-level circuits that compute hypothesized intermediate variables. This enables causal analysis through targeted ablation at the level of model parameters. We apply this method to models trained on simple arithmetic tasks, demonstrating its effectiveness at (1) deciphering the algorithms that models have learned, (2) revealing modular structure within a model, and (3) tracking the development of circuits over training. We compare circuit probing to other methods across these three experiments, and find it on par or more effective than existing analysis methods. Finally, we demonstrate circuit probing on a real-world use case, uncovering circuits that are responsible for subject-verb agreement and reflexive anaphora in GPT2-Small and Medium.
Authors: Danial Kamali, Parisa Kordjamshidi
Compositional generalization, the ability of intelligent models to extrapolate understanding of components to novel compositions, is a fundamental yet challenging facet in AI research, especially within multimodal environments. In this work, we address this challenge by exploiting the syntactic structure of language to boost compositional generalization. This paper elevates the importance of syntactic grounding, particularly through attention masking techniques derived from text input parsing. We introduce and evaluate the merits of using syntactic information in the multimodal grounding problem. Our results on grounded compositional generalization underscore the positive impact of dependency parsing across diverse tasks when utilized with Weight Sharing across the Transformer encoder. The results push the state-of-the-art in multimodal grounding and parameter-efficient modeling and provide insights for future research.
Authors: Jason Holmes, Rui Peng, Yiwei Li, Jinyu Hu, Zhengliang Liu, Zihao Wu, Huan Zhao, Xi Jiang, Wei Liu, Hong Wei, Jie Zou, Tianming Liu, Yi Shao
IMPORTANCE The response effectiveness of different large language models (LLMs) and various individuals, including medical students, graduate students, and practicing physicians, in pediatric ophthalmology consultations, has not been clearly established yet. OBJECTIVE Design a 100-question exam based on pediatric ophthalmology to evaluate the performance of LLMs in highly specialized scenarios and compare them with the performance of medical students and physicians at different levels. DESIGN, SETTING, AND PARTICIPANTS This survey study assessed three LLMs, namely ChatGPT (GPT-3.5), GPT-4, and PaLM2, were assessed alongside three human cohorts: medical students, postgraduate students, and attending physicians, in their ability to answer questions related to pediatric ophthalmology. It was conducted by administering questionnaires in the form of test papers through the LLM network interface, with the valuable participation of volunteers. MAIN OUTCOMES AND MEASURES Mean scores of LLM and humans on 100 multiple-choice questions, as well as the answer stability, correlation, and response confidence of each LLM. RESULTS GPT-4 performed comparably to attending physicians, while ChatGPT (GPT-3.5) and PaLM2 outperformed medical students but slightly trailed behind postgraduate students. Furthermore, GPT-4 exhibited greater stability and confidence when responding to inquiries compared to ChatGPT (GPT-3.5) and PaLM2. CONCLUSIONS AND RELEVANCE Our results underscore the potential for LLMs to provide medical assistance in pediatric ophthalmology and suggest significant capacity to guide the education of medical students.
Authors: Hanlin Zhang, Benjamin L. Edelman, Danilo Francati, Daniele Venturi, Giuseppe Ateniese, Boaz Barak
Watermarking generative models consists of planting a statistical signal (watermark) in a model's output so that it can be later verified that the output was generated by the given model. A strong watermarking scheme satisfies the property that a computationally bounded attacker cannot erase the watermark without causing significant quality degradation. In this paper, we study the (im)possibility of strong watermarking schemes. We prove that, under well-specified and natural assumptions, strong watermarking is impossible to achieve. This holds even in the private detection algorithm setting, where the watermark insertion and detection algorithms share a secret key, unknown to the attacker. To prove this result, we introduce a generic efficient watermark attack; the attacker is not required to know the private key of the scheme or even which scheme is used. Our attack is based on two assumptions: (1) The attacker has access to a "quality oracle" that can evaluate whether a candidate output is a high-quality response to a prompt, and (2) The attacker has access to a "perturbation oracle" which can modify an output with a nontrivial probability of maintaining quality, and which induces an efficiently mixing random walk on high-quality outputs. We argue that both assumptions can be satisfied in practice by an attacker with weaker computational capabilities than the watermarked model itself, to which the attacker has only black-box access. Furthermore, our assumptions will likely only be easier to satisfy over time as models grow in capabilities and modalities. We demonstrate the feasibility of our attack by instantiating it to attack three existing watermarking schemes for large language models: Kirchenbauer et al. (2023), Kuditipudi et al. (2023), and Zhao et al. (2023). The same attack successfully removes the watermarks planted by all three schemes, with only minor quality degradation.
Authors: Xiang Zhou, Yichen Jiang, Mohit Bansal
Recent diagnostic datasets on compositional generalization, such as SCAN (Lake and Baroni, 2018) and COGS (Kim and Linzen, 2020), expose severe problems in models trained from scratch on these datasets. However, in contrast to this poor performance, state-of-the-art models trained on larger and more general datasets show better generalization ability. In this work, to reconcile this inconsistency, we conduct an empirical analysis by training Transformer models on a variety of training sets with different data factors, including dataset scale, pattern complexity, example difficulty, etc. First, we show that increased dataset complexity can lead to better generalization behavior on multiple different generalization challenges. To further understand this improvement, we show two axes of the benefit from more complex datasets: they provide more diverse examples so compositional understanding becomes more effective, and they also prevent ungeneralizable memorization of the examples due to reduced example repetition frequency. Finally, we explore how training examples of different difficulty levels influence generalization differently. On synthetic datasets, simple examples invoke stronger compositionality than hard examples do. On larger-scale real language datasets, while hard examples become more important potentially to ensure decent data coverage, a balanced mixture of simple and hard examples manages to induce the strongest generalizability. The code and data for this work are available at https://github.com/owenzx/data4comp
Authors: Jishnu Ray Chowdhury, Cornelia Caragea
Binary Balanced Tree RvNNs (BBT-RvNNs) enforce sequence composition according to a preset balanced binary tree structure. Thus, their non-linear recursion depth is just $\log_2 n$ ($n$ being the sequence length). Such logarithmic scaling makes BBT-RvNNs efficient and scalable on long sequence tasks such as Long Range Arena (LRA). However, such computational efficiency comes at a cost because BBT-RvNNs cannot solve simple arithmetic tasks like ListOps. On the flip side, RvNNs (e.g., Beam Tree RvNN) that do succeed on ListOps (and other structure-sensitive tasks like formal logical inference) are generally several times more expensive than even RNNs. In this paper, we introduce a novel framework -- Recursion in Recursion (RIR) to strike a balance between the two sides - getting some of the benefits from both worlds. In RIR, we use a form of two-level nested recursion - where the outer recursion is a $k$-ary balanced tree model with another recursive model (inner recursion) implementing its cell function. For the inner recursion, we choose Beam Tree RvNNs (BT-RvNN). To adjust BT-RvNNs within RIR we also propose a novel strategy of beam alignment. Overall, this entails that the total recursive depth in RIR is upper-bounded by $k \log_k n$. Our best RIR-based model is the first model that demonstrates high ($\geq 90\%$) length-generalization performance on ListOps while at the same time being scalable enough to be trainable on long sequence inputs from LRA. Moreover, in terms of accuracy in the LRA language tasks, it performs competitively with Structured State Space Models (SSMs) without any special initialization - outperforming Transformers by a large margin. On the other hand, while SSMs can marginally outperform RIR on LRA, they (SSMs) fail to length-generalize on ListOps. Our code is available at: \url{https://github.com/JRC1995/BeamRecursionFamily/}.
Authors: Ryo Ueda, Tadahiro Taniguchi
As a sub-discipline of evolutionary and computational linguistics, emergent communication (EC) studies communication protocols, called emergent languages, arising in simulations where agents communicate. A key goal of EC is to give rise to languages that share statistical properties with natural languages. In this paper, we reinterpret Lewis's signaling game, a frequently used setting in EC, as beta-VAE and reformulate its objective function as ELBO. Consequently, we clarify the existence of prior distributions of emergent languages and show that the choice of the priors can influence their statistical properties. Specifically, we address the properties of word lengths and segmentation, known as Zipf's law of abbreviation (ZLA) and Harris's articulation scheme (HAS), respectively. It has been reported that the emergent languages do not follow them when using the conventional objective. We experimentally demonstrate that by selecting an appropriate prior distribution, more natural segments emerge, while suggesting that the conventional one prevents the languages from following ZLA and HAS.
Authors: Yichen Wang, Kevin Yang, Xiaoming Liu, Dan Klein
Existing LLM-based systems for writing long-form stories or story outlines frequently suffer from unnatural pacing, whether glossing over important events or over-elaborating on insignificant details, resulting in a jarring experience for the reader. We propose a CONCrete Outline ConTrol (CONCOCT) system to improve pacing when automatically generating story outlines. We first train a concreteness evaluator to judge which of two events is more concrete (low-level-detailed). This evaluator can then be used to control pacing in hierarchical outline generation; in this work, we explore a vaguest-first expansion procedure that aims for uniform pacing. We further use the evaluator to filter new outline items based on predicted concreteness. Compared to a baseline hierarchical outline generator, humans judge CONCOCT's pacing to be more consistent over 57% of the time across multiple outline lengths; the gains also translate to downstream stories. All code, data, and models are open-sourced.
Authors: Xusheng Zhao, Hao Peng, Qiong Dai, Xu Bai, Huailiang Peng, Yanbing Liu, Qinglang Guo, Philip S. Yu
Aspect-based sentiment analysis (ABSA) is dedicated to forecasting the sentiment polarity of aspect terms within sentences. Employing graph neural networks to capture structural patterns from syntactic dependency parsing has been confirmed as an effective approach for boosting ABSA. In most works, the topology of dependency trees or dependency-based attention coefficients is often loosely regarded as edges between aspects and opinions, which can result in insufficient and ambiguous syntactic utilization. To address these problems, we propose a new reinforced dependency graph convolutional network (RDGCN) that improves the importance calculation of dependencies in both distance and type views. Initially, we propose an importance calculation criterion for the minimum distances over dependency trees. Under the criterion, we design a distance-importance function that leverages reinforcement learning for weight distribution search and dissimilarity control. Since dependency types often do not have explicit syntax like tree distances, we use global attention and mask mechanisms to design type-importance functions. Finally, we merge these weights and implement feature aggregation and classification. Comprehensive experiments on three popular datasets demonstrate the effectiveness of the criterion and importance functions. RDGCN outperforms state-of-the-art GNN-based baselines in all validations.
Authors: Wenbo Zhu, Tiechuan Hu
In this paper, we look at a database of tweets sorted by various keywords that could indicate the users sentiment towards covid vaccines. With social media becoming such a prevalent source of opinion, sorting and ranking tweets that hold important information such as opinions on covid vaccines is of utmost importance. Two different ranking scales were used, and ranking a tweet in this way could represent the difference between an opinion being lost and an opinion being featured on the site, which affects the decisions and behavior of people, and why researchers were interested in it. Using natural language processing techniques, our aim is to determine and categorize opinions about covid vaccines with the highest accuracy possible.
Authors: Cheng-Yu Chuang, Pooyan Fazli
Video description entails automatically generating coherent natural language sentences that narrate the content of a given video. We introduce CLearViD, a transformer-based model for video description generation that leverages curriculum learning to accomplish this task. In particular, we investigate two curriculum strategies: (1) progressively exposing the model to more challenging samples by gradually applying a Gaussian noise to the video data, and (2) gradually reducing the capacity of the network through dropout during the training process. These methods enable the model to learn more robust and generalizable features. Moreover, CLearViD leverages the Mish activation function, which provides non-linearity and non-monotonicity and helps alleviate the issue of vanishing gradients. Our extensive experiments and ablation studies demonstrate the effectiveness of the proposed model. The results on two datasets, namely ActivityNet Captions and YouCook2, show that CLearViD significantly outperforms existing state-of-the-art models in terms of both accuracy and diversity metrics.
Authors: Zhengyuan Liu, Hai Leong Chieu, Nancy F. Chen
Data collection from manual labeling provides domain-specific and task-aligned supervision for data-driven approaches, and a critical mass of well-annotated resources is required to achieve reasonable performance in natural language processing tasks. However, manual annotations are often challenging to scale up in terms of time and budget, especially when domain knowledge, capturing subtle semantic features, and reasoning steps are needed. In this paper, we investigate the efficacy of leveraging large language models on automated labeling for computational stance detection. We empirically observe that while large language models show strong potential as an alternative to human annotators, their sensitivity to task-specific instructions and their intrinsic biases pose intriguing yet unique challenges in machine annotation. We introduce a multi-label and multi-target sampling strategy to optimize the annotation quality. Experimental results on the benchmark stance detection corpora show that our method can significantly improve performance and learning efficacy.
Authors: Ao Zhang, Liming Zhao, Chen-Wei Xie, Yun Zheng, Wei Ji, Tat-Seng Chua
The development of large language models (LLMs) has greatly advanced the field of multimodal understanding, leading to the emergence of large multimodal models (LMMs). In order to enhance the level of visual comprehension, recent studies have equipped LMMs with region-level understanding capabilities by representing object bounding box coordinates as a series of text sequences (pixel2seq). In this paper, we introduce a novel paradigm for object location modeling called pixel2emb method, where we ask the LMM to output the location embeddings and then decoded by different decoders. This paradigm allows for different location formats (such as bounding boxes and masks) to be used in multimodal conversations Furthermore, this kind of embedding based location modeling enables the utilization of existing practices in localization tasks, such as detection and segmentation. In scenarios with limited resources, our pixel2emb demonstrates superior performance compared to existing state-of-the-art (SOTA) approaches in both the location input and output tasks under fair comparison. Leveraging the proposed pixel2emb method, we train an LMM named NExT-Chat and demonstrate its capability of handling multiple tasks like visual grounding, region caption, and grounded reasoning.
Authors: Cam-Van Thi Nguyen, Anh-Tuan Mai, The-Son Le, Hai-Dang Kieu, Duc-Trong Le
Emotion recognition is a crucial task for human conversation understanding. It becomes more challenging with the notion of multimodal data, e.g., language, voice, and facial expressions. As a typical solution, the global- and the local context information are exploited to predict the emotional label for every single sentence, i.e., utterance, in the dialogue. Specifically, the global representation could be captured via modeling of cross-modal interactions at the conversation level. The local one is often inferred using the temporal information of speakers or emotional shifts, which neglects vital factors at the utterance level. Additionally, most existing approaches take fused features of multiple modalities in an unified input without leveraging modality-specific representations. Motivating from these problems, we propose the Relational Temporal Graph Neural Network with Auxiliary Cross-Modality Interaction (CORECT), an novel neural network framework that effectively captures conversation-level cross-modality interactions and utterance-level temporal dependencies with the modality-specific manner for conversation understanding. Extensive experiments demonstrate the effectiveness of CORECT via its state-of-the-art results on the IEMOCAP and CMU-MOSEI datasets for the multimodal ERC task.
Authors: Qian Chen, Wen Wang, Qinglin Zhang, Siqi Zheng, Shiliang Zhang, Chong Deng, Yukun Ma, Hai Yu, Jiaqing Liu, Chong Zhang
Recently, unified speech-text models, such as SpeechGPT, VioLA, and AudioPaLM, have achieved remarkable performance on speech tasks. These models convert continuous speech signals into discrete tokens (speech discretization) and merge text and speech tokens into a shared vocabulary. Then they train a single decoder-only Transformer on a mixture of speech tasks. Specifically, all these models utilize Loss Masking on the input speech tokens for the ASR task, which means that these models do not explicitly model the dependency between the speech tokens. In this paper, we attempt to model the sequence of speech tokens in an autoregressive manner like text. However, we find that applying the conventional cross-entropy loss on input speech tokens does not consistently improve the ASR performance over Loss Masking. Therefore, we propose a novel approach denoted Smoothed Label Distillation (SLD), which introduces a KL divergence loss with smoothed labels on the input speech tokens to effectively model speech tokens. Experiments demonstrate that our SLD approach alleviates the limitations of the cross-entropy loss and consistently outperforms Loss Masking for decoder-only Transformer based ASR using different speech discretization methods.
Authors: Tiasa Singha Roy, Priyam Basu
Research on data generation and augmentation has been focused majorly on enhancing generation models, leaving a notable gap in the exploration and refinement of methods for evaluating synthetic data. There are several text similarity metrics within the context of generated data filtering which can impact the performance of specific Natural Language Understanding (NLU) tasks, specifically focusing on intent and sentiment classification. In this study, we propose RankAug, a text-ranking approach that detects and filters out the top augmented texts in terms of being most similar in meaning with lexical and syntactical diversity. Through experiments conducted on multiple datasets, we demonstrate that the judicious selection of filtering techniques can yield a substantial improvement of up to 35% in classification accuracy for under-represented classes.
Authors: Julius Steuer, Marius Mosbach, Dietrich Klakow
Research on the cognitive plausibility of language models (LMs) has so far mostly concentrated on modelling psycholinguistic response variables such as reading times, gaze durations and N400/P600 EEG signals, while mostly leaving out the dimension of what Mahowald et al. (2023) described as formal and functional linguistic competence, and developmental plausibility. We address this gap by training a series of GPT-like language models of different sizes on the strict version of the BabyLM pretraining corpus, evaluating on the challenge tasks (BLiMP, GLUE, MSGS) and an additional reading time prediction task. We find a positive correlation between LM size and performance on all three challenge tasks, with different preferences for model width and depth in each of the tasks. In contrast, a negative correlation was found between LM size and reading time fit of linear mixed-effects models using LM surprisal as a predictor, with the second-smallest LM achieving the largest log-likelihood reduction over a baseline model without surprisal. This suggests that modelling processing effort and linguistic competence may require an approach different from training GPT-like LMs on a developmentally plausible corpus.
Authors: Vatsal Raina, Adian Liusie, Mark Gales
Multiple-choice tests are a common approach for assessing candidates' comprehension skills. Standard multiple-choice reading comprehension exams require candidates to select the correct answer option from a discrete set based on a question in relation to a contextual passage. For appropriate assessment, the distractor answer options must by definition be incorrect but plausible and diverse. However, generating good quality distractors satisfying these criteria is a challenging task for content creators. We propose automated assessment metrics for the quality of distractors in multiple-choice reading comprehension tests. Specifically, we define quality in terms of the incorrectness, plausibility and diversity of the distractor options. We assess incorrectness using the classification ability of a binary multiple-choice reading comprehension system. Plausibility is assessed by considering the distractor confidence - the probability mass associated with the distractor options for a standard multi-class multiple-choice reading comprehension system. Diversity is assessed by pairwise comparison of an embedding-based equivalence metric between the distractors of a question. To further validate the plausibility metric we compare against candidate distributions over multiple-choice questions and agreement with a ChatGPT model's interpretation of distractor plausibility and diversity.
Authors: Urban Knupleš, Diego Frassinelli, Sabine Schulte im Walde
Humans tend to strongly agree on ratings on a scale for extreme cases (e.g., a CAT is judged as very concrete), but judgements on mid-scale words exhibit more disagreement. Yet, collected rating norms are heavily exploited across disciplines. Our study focuses on concreteness ratings and (i) implements correlations and supervised classification to identify salient multi-modal characteristics of mid-scale words, and (ii) applies a hard clustering to identify patterns of systematic disagreement across raters. Our results suggest to either fine-tune or filter mid-scale target words before utilising them.
Authors: Zhen Yang, Yingxue Zhang, Fandong Meng, Jie Zhou
Despite Multi-modal Large Language Models (MM-LLMs) have made exciting strides recently, they are still struggling to efficiently model the interactions among multi-modal inputs and the generation in non-textual modalities. In this work, we propose TEAL (Tokenize and Embed ALl)}, an approach to treat the input from any modality as a token sequence and learn a joint embedding space for all modalities. Specifically, for the input from any modality, TEAL first discretizes it into a token sequence with the off-the-shelf tokenizer and embeds the token sequence into a joint embedding space with a learnable embedding matrix. MM-LLMs just need to predict the multi-modal tokens autoregressively as the textual LLMs do. Finally, the corresponding de-tokenizer is applied to generate the output in each modality based on the predicted token sequence. With the joint embedding space, TEAL enables the frozen LLMs to perform both understanding and generation tasks involving non-textual modalities, such as image and audio. Thus, the textual LLM can just work as an interface and maintain its high performance in textual understanding and generation. Experiments show that TEAL achieves substantial improvements in multi-modal understanding, and implements a simple scheme for multi-modal generations.
Authors: Chenmien Tan, Ge Zhang, Jie Fu
While large language models (LLMs) have enabled learning knowledge from the pre-training corpora, the acquired knowledge may be fundamentally incorrect or outdated over time, which necessitates rectifying the knowledge of the language model (LM) after the training. A promising approach involves employing a hyper-network to generate parameter shift, whereas existing hyper-networks suffer from inferior scalability in synchronous editing operation amount. To mitigate the problem, we propose the MAssive Language Model Editing Network (MALMEN), which formulates the parameter shift aggregation as the least square problem, subsequently updating the LM parameters using the normal equation. To accommodate editing multiple facts simultaneously with limited memory budgets, we separate the computation on the hyper-network and LM, enabling arbitrary batch size on both neural networks. Our method is evaluated by editing up to thousands of facts on LMs with different architectures, i.e., BERT-base, GPT-2, T5-XL (2.8B), and GPT-J (6B), across various knowledge-intensive NLP tasks, i.e., closed book fact-checking and question answering. Remarkably, MALMEN is capable of editing hundreds of times more facts than strong baselines with the identical hyper-network architecture and outperforms editor specifically designed for GPT. Our code is available at https://github.com/ChenmienTan/malmen.
Authors: Subba Reddy Oota, Emin Çelik, Fatma Deniz, Mariya Toneva
Despite known differences between reading and listening in the brain, recent work has shown that text-based language models predict both text-evoked and speech-evoked brain activity to an impressive degree. This poses the question of what types of information language models truly predict in the brain. We investigate this question via a direct approach, in which we eliminate information related to specific low-level stimulus features (textual, speech, and visual) in the language model representations, and observe how this intervention affects the alignment with fMRI brain recordings acquired while participants read versus listened to the same naturalistic stories. We further contrast our findings with speech-based language models, which would be expected to predict speech-evoked brain activity better, provided they model language processing in the brain well. Using our direct approach, we find that both text-based and speech-based language models align well with early sensory regions due to shared low-level features. Text-based models continue to align well with later language regions even after removing these features, while, surprisingly, speech-based models lose most of their alignment. These findings suggest that speech-based models can be further improved to better reflect brain-like language processing.
Authors: Khushi Bhardwaj, Raj Sanjay Shah, Sashank Varma
Pre-trained Large Language Models (LLMs) have shown success in a diverse set of language inference and understanding tasks. The pre-training stage of LLMs looks at a large corpus of raw textual data. The BabyLM shared task compares LLM pre-training to human language acquisition, where the number of tokens seen by 13-year-old kids is magnitudes smaller than the number of tokens seen by LLMs. In this work, we pre-train and evaluate LLMs on their ability to learn contextual word representations using roughly the same number of tokens as seen by children. We provide a strong set of baselines; with different architectures, evaluation of changes in performance across epochs, and reported pre-training metrics for the strict small and strict tracks of the task. We also try to loosely replicate the RoBERTa baseline given by the task organizers to observe the training robustness to hyperparameter selection and replicability. We provide the submission details to the strict and strict-small tracks in this report.
Authors: Lukas Gienapp, Harrisen Scells, Niklas Deckers, Janek Bevendorff, Shuai Wang, Johannes Kiesel, Shahbaz Syed, Maik Fröbe, Guide Zucoon, Benno Stein, Matthias Hagen, Martin Potthast
Recent advances in large language models have enabled the development of viable generative information retrieval systems. A generative retrieval system returns a grounded generated text in response to an information need instead of the traditional document ranking. Quantifying the utility of these types of responses is essential for evaluating generative retrieval systems. As the established evaluation methodology for ranking-based ad hoc retrieval may seem unsuitable for generative retrieval, new approaches for reliable, repeatable, and reproducible experimentation are required. In this paper, we survey the relevant information retrieval and natural language processing literature, identify search tasks and system architectures in generative retrieval, develop a corresponding user model, and study its operationalization. This theoretical analysis provides a foundation and new insights for the evaluation of generative ad hoc retrieval systems.
Authors: Antonios Georgiou Tankut Can, Mikhail Katkov, Misha Tsodyks
One of the most impressive achievements of the AI revolution is the development of large language models that can generate meaningful text and respond to instructions in plain English with no additional training necessary. Here we show that language models can be used as a scientific instrument for studying human memory for meaningful material. We developed a pipeline for designing large scale memory experiments and analyzing the obtained results. We performed online memory experiments with a large number of participants and collected recognition and recall data for narratives of different lengths. We found that both recall and recognition performance scale linearly with narrative length. Furthermore, in order to investigate the role of narrative comprehension in memory, we repeated these experiments using scrambled versions of the presented stories. We found that even though recall performance declined significantly, recognition remained largely unaffected. Interestingly, recalls in this condition seem to follow the original narrative order rather than the scrambled presentation, pointing to a contextual reconstruction of the story in memory.
Authors: Md Azim Khan
Online conversations can be toxic and subjected to threats, abuse, or harassment. To identify toxic text comments, several deep learning and machine learning models have been proposed throughout the years. However, recent studies demonstrate that because of the imbalances in the training data, some models are more likely to show unintended biases including gender bias and identity bias. In this research, our aim is to detect toxic comment and reduce the unintended bias concerning identity features such as race, gender, sex, religion by fine-tuning an attention based model called BERT(Bidirectional Encoder Representation from Transformers). We apply weighted loss to address the issue of unbalanced data and compare the performance of a fine-tuned BERT model with a traditional Logistic Regression model in terms of classification and bias minimization. The Logistic Regression model with the TFIDF vectorizer achieve 57.1% accuracy, and fine-tuned BERT model's accuracy is 89%. Code is available at https://github.com/zim10/Determine_Toxic_comment_and_identity_bias.git
Authors: Martin Kuo, Jianyi Zhang, Yiran Chen
Building on the cost-efficient pretraining advancements brought about by Crammed BERT, we enhance its performance and interpretability further by introducing a novel pretrained model Dependency Agreement Crammed BERT (DACBERT) and its two-stage pretraining framework - Dependency Agreement Pretraining. This framework, grounded by linguistic theories, seamlessly weaves syntax and semantic information into the pretraining process. The first stage employs four dedicated submodels to capture representative dependency agreements at the chunk level, effectively converting these agreements into embeddings. The second stage uses these refined embeddings, in tandem with conventional BERT embeddings, to guide the pretraining of the rest of the model. Evaluated on the GLUE benchmark, our DACBERT demonstrates notable improvement across various tasks, surpassing Crammed BERT by 3.13% in the RTE task and by 2.26% in the MRPC task. Furthermore, our method boosts the average GLUE score by 0.83%, underscoring its significant potential. The pretraining process can be efficiently executed on a single GPU within a 24-hour cycle, necessitating no supplementary computational resources or extending the pretraining duration compared with the Crammed BERT. Extensive studies further illuminate our approach's instrumental role in bolstering the interpretability of pretrained language models for natural language understanding tasks.
Authors: Zheng Chu, Zekun Wang, Jiafeng Liang, Ming Liu, Bing Qin
The facts and time in the document are intricately intertwined, making temporal reasoning over documents challenging. Previous work models time implicitly, making it difficult to handle such complex relationships. To address this issue, we propose MTGER, a novel Multi-view Temporal Graph Enhanced Temporal Reasoning framework for temporal reasoning over time-involved documents. Concretely, MTGER explicitly models the temporal relationships among facts by multi-view temporal graphs. On the one hand, the heterogeneous temporal graphs explicitly model the temporal and discourse relationships among facts; on the other hand, the multi-view mechanism captures both time-focused and fact-focused information, allowing the two views to complement each other through adaptive fusion. To further improve the implicit reasoning capability of the model, we design a self-supervised time-comparing objective. Extensive experimental results demonstrate the effectiveness of our method on the TimeQA and SituatedQA datasets. Furthermore, MTGER gives more consistent answers under question perturbations.
Authors: Zhen Qin, Songlin Yang, Yiran Zhong
Transformers have surpassed RNNs in popularity due to their superior abilities in parallel training and long-term dependency modeling. Recently, there has been a renewed interest in using linear RNNs for efficient sequence modeling. These linear RNNs often employ gating mechanisms in the output of the linear recurrence layer while ignoring the significance of using forget gates within the recurrence. In this paper, we propose a gated linear RNN model dubbed Hierarchically Gated Recurrent Neural Network (HGRN), which includes forget gates that are lower bounded by a learnable value. The lower bound increases monotonically when moving up layers. This allows the upper layers to model long-term dependencies and the lower layers to model more local, short-term dependencies. Experiments on language modeling, image classification, and long-range arena benchmarks showcase the efficiency and effectiveness of our proposed model. The source code is available at https://github.com/OpenNLPLab/HGRN.
Authors: Shuo Yang, Wei-Lin Chiang, Lianmin Zheng, Joseph E. Gonzalez, Ion Stoica
Large language models are increasingly trained on all the data ever produced by humans. Many have raised concerns about the trustworthiness of public benchmarks due to potential contamination in pre-training or fine-tuning datasets. While most data decontamination efforts apply string matching (e.g., n-gram overlap) to remove benchmark data, we show that these methods are insufficient, and simple variations of test data (e.g., paraphrasing, translation) can easily bypass these decontamination measures. Furthermore, we demonstrate that if such variation of test data is not eliminated, a 13B model can easily overfit a test benchmark and achieve drastically high performance, on par with GPT-4. We validate such observations in widely used benchmarks such as MMLU, GSK8k, and HumanEval. To address this growing risk, we propose a stronger LLM-based decontamination method and apply it to widely used pre-training and fine-tuning datasets, revealing significant previously unknown test overlap. For example, in pre-training sets such as RedPajama-Data-1T and StarCoder-Data, we identified that 8-18\% of the HumanEval benchmark overlaps. Interestingly, we also find such contamination in synthetic dataset generated by GPT-3.5/4, suggesting a potential risk of unintentional contamination. We urge the community to adopt stronger decontamination approaches when using public benchmarks. Moreover, we call for the community to actively develop fresh one-time exams to evaluate models accurately. Our decontamination tool is publicly available at https://github.com/lm-sys/llm-decontaminator.
Authors: Jianxin Yang
We present LongQLoRA, an efficient and effective method to extend context length of large language models with less training resources. LongQLoRA combines the advantages of Position Interpolation, QLoRA and Shift Short Attention of LongLoRA. With a single 32GB V100 GPU, LongQLoRA can extend the context length of LLaMA2 7B and 13B from 4096 to 8192 and even to 12k within 1000 finetuning steps. LongQLoRA achieves competitive perplexity performance on PG19 and Proof-pile datasets, our model outperforms LongLoRA and is very close to MPT-7B-8K within the evaluation context length of 8192. We collect and build 39k long instruction data to extend context length of Vicuna-13B from 4096 to 8192 and achieve good performance both in long and short context generation task. We also do some ablation experiments to study the effect of LoRA rank, finetuning steps and attention patterns in inference.The model weights, training data and code are avaliable at https://github.com/yangjianxin1/LongQLoRA.
Authors: Tibor L. R. Krols, Marie Mortensen, Ninell Oldenburg
Social media has become a very popular source of information. With this popularity comes an interest in systems that can classify the information produced. This study tries to create such a system detecting irony in Twitter users. Recent work emphasize the importance of lexical features, sentiment features and the contrast herein along with TF-IDF and topic models. Based on a thorough feature selection process, the resulting model contains specific sub-features from these areas. Our model reaches an F1-score of 0.84, which is above the baseline. We find that lexical features, especially TF-IDF, contribute the most to our models while sentiment and topic modeling features contribute less to overall performance. Lastly, we highlight multiple interesting and important paths for further exploration.
Authors: Tal Schuster, Adam D. Lelkes, Haitian Sun, Jai Gupta, Jonathan Berant, William W. Cohen, Donald Metzler
Recently proposed long-form question answering (QA) systems, supported by large language models (LLMs), have shown promising capabilities. Yet, attributing and verifying their generated abstractive answers can be difficult, and automatically evaluating their accuracy remains an ongoing challenge.
In this work, we introduce a new QA task for answering multi-answer questions by summarizing multiple diverse sources in a semi-extractive fashion. Specifically, Semi-extractive Multi-source QA (SEMQA) requires models to output a comprehensive answer, while mixing factual quoted spans -- copied verbatim from given input sources -- and non-factual free-text connectors that glue these spans together into a single cohesive passage. This setting bridges the gap between the outputs of well-grounded but constrained extractive QA systems and more fluent but harder to attribute fully abstractive answers. Particularly, it enables a new mode for language models that leverages their advanced language generation capabilities, while also producing fine in-line attributions by-design that are easy to verify, interpret, and evaluate.
To study this task, we create the first dataset of this kind, QuoteSum, with human-written semi-extractive answers to natural and generated questions, and define text-based evaluation metrics. Experimenting with several LLMs in various settings, we find this task to be surprisingly challenging, demonstrating the importance of QuoteSum for developing and studying such consolidation capabilities.
Authors: Shashank Gupta, Vaishnavi Shrivastava, Ameet Deshpande, Ashwin Kalyan, Peter Clark, Ashish Sabharwal, Tushar Khot
Recent works have showcased the ability of large-scale language models (LLMs) to embody diverse personas in their responses, exemplified by prompts like 'You are Yoda. Explain the Theory of Relativity.' While this ability allows personalization of LLMs and enables human behavior simulation, its effect on LLMs' capabilities remain unclear. To fill this gap, we present the first extensive study of the unintended side-effects of persona assignment on the ability of LLMs, specifically ChatGPT, to perform basic reasoning tasks. Our study covers 24 reasoning datasets and 16 diverse personas spanning 5 socio-demographic groups: race, gender, religion, disability, and political affiliation. Our experiments unveil that ChatGPT carries deep rooted bias against various socio-demographics underneath a veneer of fairness. While it overtly rejects stereotypes when explicitly asked ('Are Black people less skilled at mathematics?'), it manifests stereotypical and often erroneous presumptions when prompted to answer questions while taking on a persona. These can be observed as abstentions in the model responses, e.g., 'As a Black person, I am unable to answer this question as it requires math knowledge', and generally result in a substantial drop in performance on reasoning tasks. We find that this inherent deep bias is ubiquitous - 80% of our personas demonstrated bias; it is significant - certain datasets had relative drops in performance of 70%+; and can be especially harmful for certain groups - certain personas had stat. sign. drops on more than 80% of the datasets. Further analysis shows that these persona-induced errors can be hard-to-discern and hard-to-avoid. Our findings serve as a cautionary tale that the practice of assigning personas to LLMs - a trend on the rise - can surface their deep-rooted biases and have unforeseeable and detrimental side-effects.
Authors: Koyena Pal, Jiuding Sun, Andrew Yuan, Byron C. Wallace, David Bau
We conjecture that hidden state vectors corresponding to individual input tokens encode information sufficient to accurately predict several tokens ahead. More concretely, in this paper we ask: Given a hidden (internal) representation of a single token at position $t$ in an input, can we reliably anticipate the tokens that will appear at positions $\geq t + 2$? To test this, we measure linear approximation and causal intervention methods in GPT-J-6B to evaluate the degree to which individual hidden states in the network contain signal rich enough to predict future hidden states and, ultimately, token outputs. We find that, at some layers, we can approximate a model's output with more than 48% accuracy with respect to its prediction of subsequent tokens through a single hidden state. Finally we present a "Future Lens" visualization that uses these methods to create a new view of transformer states.
Authors: Michael Wilson, Jackson Petty, Robert Frank
Language models are typically evaluated on their success at predicting the distribution of specific words in specific contexts. Yet linguistic knowledge also encodes relationships between contexts, allowing inferences between word distributions. We investigate the degree to which pre-trained Transformer-based large language models (LLMs) represent such relationships, focusing on the domain of argument structure. We find that LLMs perform well in generalizing the distribution of a novel noun argument between related contexts that were seen during pre-training (e.g., the active object and passive subject of the verb spray), succeeding by making use of the semantically-organized structure of the embedding space for word embeddings. However, LLMs fail at generalizations between related contexts that have not been observed during pre-training, but which instantiate more abstract, but well-attested structural generalizations (e.g., between the active object and passive subject of an arbitrary verb). Instead, in this case, LLMs show a bias to generalize based on linear order. This finding points to a limitation with current models and points to a reason for which their training is data-intensive.s reported here are available at https://github.com/clay-lab/structural-alternations.
Authors: Rocktim Jyoti Das, Liqun Ma, Zhiqiang Shen
Large Language Models (LLMs) with a billion or more parameters are prime targets for network pruning, which aims to reduce a portion of the network weights without compromising performance. Prior approaches such as Weights Magnitude, SparseGPT, and Wanda, either concentrated solely on weights or integrated weights with activations for sparsity. However, they overlooked the informative gradients derived from pretrained large language models. In this paper, we present a novel sparsity-centric pruning method for pretrained LLMs, termed Gradient-based Language Model Pruner (GBLM-Pruner). GBLM-Pruner leverages the first-order term of the Taylor expansion, operating in a training-free manner by harnessing properly normalized gradients from a few calibration samples to determine the importance pruning score, and substantially outperforms competitive counterparts like SparseGPT and Wanda in multiple benchmarks. Intriguing, after incorporating gradients, the unstructured pruning method tends to reveal some structural patterns post-pruning, which mirrors the geometric interdependence inherent in the LLMs' parameter structure. Additionally, GBLM-Pruner functions without any subsequent retraining or weight updates to maintain its simplicity as other counterparts. Extensive evaluations on LLaMA-1 and LLaMA-2 across various language benchmarks and perplexity show that GBLM-Pruner surpasses magnitude pruning, Wanda (weights+activations) and SparseGPT (weights+activations+weight update) by significant margins. Our code and models are available at https://github.com/RocktimJyotiDas/GBLM-Pruner.
Authors: Jianlin Su, Yu Lu, Shengfeng Pan, Ahmed Murtadha, Bo Wen, Yunfeng Liu
Position encoding recently has shown effective in the transformer architecture. It enables valuable supervision for dependency modeling between elements at different positions of the sequence. In this paper, we first investigate various methods to integrate positional information into the learning process of transformer-based language models. Then, we propose a novel method named Rotary Position Embedding(RoPE) to effectively leverage the positional information. Specifically, the proposed RoPE encodes the absolute position with a rotation matrix and meanwhile incorporates the explicit relative position dependency in self-attention formulation. Notably, RoPE enables valuable properties, including the flexibility of sequence length, decaying inter-token dependency with increasing relative distances, and the capability of equipping the linear self-attention with relative position encoding. Finally, we evaluate the enhanced transformer with rotary position embedding, also called RoFormer, on various long text classification benchmark datasets. Our experiments show that it consistently overcomes its alternatives. Furthermore, we provide a theoretical analysis to explain some experimental results. RoFormer is already integrated into Huggingface: \url{https://huggingface.co/docs/transformers/model_doc/roformer}.
Authors: Julia Rozanova, Deborah Ferreira, Marco Valentino, Mokanrarangan Thayaparan, Andre Freitas
In the interest of interpreting neural NLI models and their reasoning strategies, we carry out a systematic probing study which investigates whether these models capture the crucial semantic features central to natural logic: monotonicity and concept inclusion. Correctly identifying valid inferences in downward-monotone contexts is a known stumbling block for NLI performance, subsuming linguistic phenomena such as negation scope and generalized quantifiers. To understand this difficulty, we emphasize monotonicity as a property of a context and examine the extent to which models capture monotonicity information in the contextual embeddings which are intermediate to their decision making process. Drawing on the recent advancement of the probing paradigm, we compare the presence of monotonicity features across various models. We find that monotonicity information is notably weak in the representations of popular NLI models which achieve high scores on benchmarks, and observe that previous improvements to these models based on fine-tuning strategies have introduced stronger monotonicity features together with their improved performance on challenge sets.
Authors: William Chen, Siyi Hu, Rajat Talak, Luca Carlone
Abstract semantic 3D scene understanding is a problem of critical importance in robotics. As robots still lack the common-sense knowledge about household objects and locations of an average human, we investigate the use of pre-trained language models to impart common sense for scene understanding. We introduce and compare a wide range of scene classification paradigms that leverage language only (zero-shot, embedding-based, and structured-language) or vision and language (zero-shot and fine-tuned). We find that the best approaches in both categories yield $\sim 70\%$ room classification accuracy, exceeding the performance of pure-vision and graph classifiers. We also find such methods demonstrate notable generalization and transfer capabilities stemming from their use of language.
Authors: Xing Han, Tongzheng Ren, Tan Minh Nguyen, Khai Nguyen, Joydeep Ghosh, Nhat Ho
Recent advances in Transformer architectures have empowered their empirical success in a variety of tasks across different domains. However, existing works mainly focus on predictive accuracy and computational cost, without considering other practical issues, such as robustness to contaminated samples. Recent work by Nguyen et al., (2022) has shown that the self-attention mechanism, which is the center of the Transformer architecture, can be viewed as a non-parametric estimator based on kernel density estimation (KDE). This motivates us to leverage a set of robust kernel density estimation methods for alleviating the issue of data contamination. Specifically, we introduce a series of self-attention mechanisms that can be incorporated into different Transformer architectures and discuss the special properties of each method. We then perform extensive empirical studies on language modeling and image classification tasks. Our methods demonstrate robust performance in multiple scenarios while maintaining competitive results on clean datasets.
Authors: Christo Kurisummoottil Thomas, Walid Saad
Semantic communication (SC) aims to communicate reliably with minimal data transfer while simultaneously providing seamless connectivity to heterogeneous services and users. In this paper, a novel emergent SC (ESC) system framework is proposed and is composed of a signaling game for emergent language design and a neuro-symbolic (NeSy) artificial intelligence (AI) approach for causal reasoning. In order to design the language, the signaling game is solved using an alternating maximization between the communicating node's utilities. The emergent language helps create a context-aware transmit vocabulary (minimal semantic representation) and aids the reasoning process (enabling generalization to unseen scenarios) by splitting complex messages into simpler reasoning tasks for the receiver. The causal description at the transmitter is then modeled (a neural component) as a posterior distribution of the relevant attributes present in the data. Using the reconstructed causal state, the receiver evaluates a set of logical formulas (symbolic part) to execute its task. The nodes NeSy reasoning components are implemented by the recently proposed AI tool called Generative Flow Networks, and they are optimized for higher semantic reliability. The ESC system is designed to enhance the novel metrics of semantic information, reliability, distortion and similarity that are designed using rigorous algebraic properties from category theory thereby generalizing the metrics beyond Shannon's notion of uncertainty. Simulation results validate the ability of ESC to communicate efficiently (with reduced bits) and achieve better semantic reliability than conventional wireless and state-of-the-art systems that do not exploit causal reasoning capabilities.
Authors: Subba Reddy Oota, Manish Gupta, Mariya Toneva
Language models have been shown to be very effective in predicting brain recordings of subjects experiencing complex language stimuli. For a deeper understanding of this alignment, it is important to understand the correspondence between the detailed processing of linguistic information by the human brain versus language models. We investigate this correspondence via a direct approach, in which we eliminate information related to specific linguistic properties in the language model representations and observe how this intervention affects the alignment with fMRI brain recordings obtained while participants listened to a story. We investigate a range of linguistic properties (surface, syntactic, and semantic) and find that the elimination of each one results in a significant decrease in brain alignment. Specifically, we find that syntactic properties (i.e. Top Constituents and Tree Depth) have the largest effect on the trend of brain alignment across model layers. These findings provide clear evidence for the role of specific linguistic information in the alignment between brain and language models, and open new avenues for mapping the joint information processing in both systems. We make the code publicly available [https://github.com/subbareddy248/linguistic-properties-brain-alignment].
Authors: Shyam Sudhakaran, Miguel González-Duque, Claire Glanois, Matthias Freiberger, Elias Najarro, Sebastian Risi
Procedural Content Generation (PCG) is a technique to generate complex and diverse environments in an automated way. However, while generating content with PCG methods is often straightforward, generating meaningful content that reflects specific intentions and constraints remains challenging. Furthermore, many PCG algorithms lack the ability to generate content in an open-ended manner. Recently, Large Language Models (LLMs) have shown to be incredibly effective in many diverse domains. These trained LLMs can be fine-tuned, re-using information and accelerating training for new tasks. Here, we introduce MarioGPT, a fine-tuned GPT2 model trained to generate tile-based game levels, in our case Super Mario Bros levels. MarioGPT can not only generate diverse levels, but can be text-prompted for controllable level generation, addressing one of the key challenges of current PCG techniques. As far as we know, MarioGPT is the first text-to-level model and combined with novelty search it enables the generation of diverse levels with varying play-style dynamics (i.e. player paths) and the open-ended discovery of an increasingly diverse range of content. Code available at https://github.com/shyamsn97/mario-gpt.
Authors: Susung Hong, Donghoon Ahn, Seungryong Kim
Existing score-distilling text-to-3D generation techniques, despite their considerable promise, often encounter the view inconsistency problem. One of the most notable issues is the Janus problem, where the most canonical view of an object (\textit{e.g}., face or head) appears in other views. In this work, we explore existing frameworks for score-distilling text-to-3D generation and identify the main causes of the view inconsistency problem -- the embedded bias of 2D diffusion models. Based on these findings, we propose two approaches to debias the score-distillation frameworks for view-consistent text-to-3D generation. Our first approach, called score debiasing, involves cutting off the score estimated by 2D diffusion models and gradually increasing the truncation value throughout the optimization process. Our second approach, called prompt debiasing, identifies conflicting words between user prompts and view prompts using a language model, and adjusts the discrepancy between view prompts and the viewing direction of an object. Our experimental results show that our methods improve the realism of the generated 3D objects by significantly reducing artifacts and achieve a good trade-off between faithfulness to the 2D diffusion models and 3D consistency with little overhead. Our project page is available at~\url{https://susunghong.github.io/Debiased-Score-Distillation-Sampling/}.
Authors: Michael Schlichtkrull, Nedjma Ousidhoum, Andreas Vlachos
Automated fact-checking is often presented as an epistemic tool that fact-checkers, social media consumers, and other stakeholders can use to fight misinformation. Nevertheless, few papers thoroughly discuss how. We document this by analysing 100 highly-cited papers, and annotating epistemic elements related to intended use, i.e., means, ends, and stakeholders. We find that narratives leaving out some of these aspects are common, that many papers propose inconsistent means and ends, and that the feasibility of suggested strategies rarely has empirical backing. We argue that this vagueness actively hinders the technology from reaching its goals, as it encourages overclaiming, limits criticism, and prevents stakeholder feedback. Accordingly, we provide several recommendations for thinking and writing about the use of fact-checking artefacts.
Authors: Eliza Kosoy, Emily Rose Reagan, Leslie Lai, Alison Gopnik, Danielle Krettek Cobb
Developmental psychologists have spent decades devising experiments to test the intelligence and knowledge of infants and children, tracing the origin of crucial concepts and capacities. Moreover, experimental techniques in developmental psychology have been carefully designed to discriminate the cognitive capacities that underlie particular behaviors. We propose that using classical experiments from child development is a particularly effective way to probe the computational abilities of AI models, in general, and LLMs in particular. First, the methodological techniques of developmental psychology, such as the use of novel stimuli to control for past experience or control conditions to determine whether children are using simple associations, can be equally helpful for assessing the capacities of LLMs. In parallel, testing LLMs in this way can tell us whether the information that is encoded in text is sufficient to enable particular responses, or whether those responses depend on other kinds of information, such as information from exploration of the physical world. In this work we adapt classical developmental experiments to evaluate the capabilities of LaMDA, a large language model from Google. We propose a novel LLM Response Score (LRS) metric which can be used to evaluate other language models, such as GPT. We find that LaMDA generates appropriate responses that are similar to those of children in experiments involving social understanding, perhaps providing evidence that knowledge of these domains is discovered through language. On the other hand, LaMDA's responses in early object and action understanding, theory of mind, and especially causal reasoning tasks are very different from those of young children, perhaps showing that these domains require more real-world, self-initiated exploration and cannot simply be learned from patterns in language input.
Authors: Darshan Deshpande, Zhivar Sourati, Filip Ilievski, Fred Morstatter
Automatic assessment of the quality of arguments has been recognized as a challenging task with significant implications for misinformation and targeted speech. While real-world arguments are tightly anchored in context, existing computational methods analyze their quality in isolation, which affects their accuracy and generalizability. We propose SPARK: a novel method for scoring argument quality based on contextualization via relevant knowledge. We devise four augmentations that leverage large language models to provide feedback, infer hidden assumptions, supply a similar-quality argument, or give a counter-argument. SPARK uses a dual-encoder Transformer architecture to enable the original argument and its augmentation to be considered jointly. Our experiments in both in-domain and zero-shot setups show that SPARK consistently outperforms existing techniques across multiple metrics.
Authors: Gaurav Maheshwari, Aurélien Bellet, Pascal Denis, Mikaela Keller
In this work, we consider the problem of intersectional group fairness in the classification setting, where the objective is to learn discrimination-free models in the presence of several intersecting sensitive groups. First, we illustrate various shortcomings of existing fairness measures commonly used to capture intersectional fairness. Then, we propose a new definition called the $\alpha$-Intersectional Fairness, which combines the absolute and the relative performance across sensitive groups and can be seen as a generalization of the notion of differential fairness. We highlight several desirable properties of the proposed definition and analyze its relation to other fairness measures. Finally, we benchmark multiple popular in-processing fair machine learning approaches using our new fairness definition and show that they do not achieve any improvement over a simple baseline. Our results reveal that the increase in fairness measured by previous definitions hides a "leveling down" effect, i.e., degrading the best performance over groups rather than improving the worst one.
Authors: Michael Schlichtkrull, Zhijiang Guo, Andreas Vlachos
Existing datasets for automated fact-checking have substantial limitations, such as relying on artificial claims, lacking annotations for evidence and intermediate reasoning, or including evidence published after the claim. In this paper we introduce AVeriTeC, a new dataset of 4,568 real-world claims covering fact-checks by 50 different organizations. Each claim is annotated with question-answer pairs supported by evidence available online, as well as textual justifications explaining how the evidence combines to produce a verdict. Through a multi-round annotation process, we avoid common pitfalls including context dependence, evidence insufficiency, and temporal leakage, and reach a substantial inter-annotator agreement of $\kappa=0.619$ on verdicts. We develop a baseline as well as an evaluation scheme for verifying claims through several question-answering steps against the open web.
Authors: Yaoting Wang, Yuanchao Li, Paul Pu Liang, Louis-Philippe Morency, Peter Bell, Catherine Lai
Fusing multiple modalities has proven effective for multimodal information processing. However, the incongruity between modalities poses a challenge for multimodal fusion, especially in affect recognition. In this study, we first analyze how the salient affective information in one modality can be affected by the other, and demonstrate that inter-modal incongruity exists latently in crossmodal attention. Based on this finding, we propose the Hierarchical Crossmodal Transformer with Dynamic Modality Gating (HCT-DMG), a lightweight incongruity-aware model, which dynamically chooses the primary modality in each training batch and reduces fusion times by leveraging the learned hierarchy in the latent space to alleviate incongruity. The experimental evaluation on five benchmark datasets: CMU-MOSI, CMU-MOSEI, and IEMOCAP (sentiment and emotion), where incongruity implicitly lies in hard samples, as well as UR-FUNNY (humour) and MUStaRD (sarcasm), where incongruity is common, verifies the efficacy of our approach, showing that HCT-DMG: 1) outperforms previous multimodal models with a reduced size of approximately 0.8M parameters; 2) recognizes hard samples where incongruity makes affect recognition difficult; 3) mitigates the incongruity at the latent level in crossmodal attention.
Authors: Pavan Kalyan Reddy Neerudu, Subba Reddy Oota, Mounika Marreddy, Venkateswara Rao Kagita, Manish Gupta
Transformer-based pretrained models like BERT, GPT-2 and T5 have been finetuned for a large number of natural language processing (NLP) tasks, and have been shown to be very effective. However, while finetuning, what changes across layers in these models with respect to pretrained checkpoints is under-studied. Further, how robust are these models to perturbations in input text? Does the robustness vary depending on the NLP task for which the models have been finetuned? While there exists some work on studying the robustness of BERT finetuned for a few NLP tasks, there is no rigorous study that compares this robustness across encoder only, decoder only and encoder-decoder models. In this paper, we characterize changes between pretrained and finetuned language model representations across layers using two metrics: CKA and STIR. Further, we study the robustness of three language models (BERT, GPT-2 and T5) with eight different text perturbations on classification tasks from the General Language Understanding Evaluation (GLUE) benchmark, and generation tasks like summarization, free-form generation and question generation. GPT-2 representations are more robust than BERT and T5 across multiple types of input perturbation. Although models exhibit good robustness broadly, dropping nouns, verbs or changing characters are the most impactful. Overall, this study provides valuable insights into perturbation-specific weaknesses of popular Transformer-based models, which should be kept in mind when passing inputs. We make the code and models publicly available [https://github.com/PavanNeerudu/Robustness-of-Transformers-models].
Authors: Dung Thai, Dhruv Agarwal, Mudit Chaudhary, Wenlong Zhao, Rajarshi Das, Manzil Zaheer, Jay-Yoon Lee, Hannaneh Hajishirzi, Andrew McCallum
We present an accurate and interpretable method for answer extraction in machine reading comprehension that is reminiscent of case-based reasoning (CBR) from classical AI. Our method (CBR-MRC) builds upon the hypothesis that contextualized answers to similar questions share semantic similarities with each other. Given a test question, CBR-MRC first retrieves a set of similar cases from a nonparametric memory and then predicts an answer by selecting the span in the test context that is most similar to the contextualized representations of answers in the retrieved cases. The semi-parametric nature of our approach allows it to attribute a prediction to the specific set of evidence cases, making it a desirable choice for building reliable and debuggable QA systems. We show that CBR-MRC provides high accuracy comparable with large reader models and outperforms baselines by 11.5 and 8.4 EM on NaturalQuestions and NewsQA, respectively. Further, we demonstrate the ability of CBR-MRC in identifying not just the correct answer tokens but also the span with the most relevant supporting evidence. Lastly, we observe that contexts for certain question types show higher lexical diversity than others and find that CBR-MRC is robust to these variations while performance using fully-parametric methods drops.
Authors: Yingcong Li, Kartik Sreenivasan, Angeliki Giannou, Dimitris Papailiopoulos, Samet Oymak
Chain-of-thought (CoT) is a method that enables language models to handle complex reasoning tasks by decomposing them into simpler steps. Despite its success, the underlying mechanics of CoT are not yet fully understood. In an attempt to shed light on this, our study investigates the impact of CoT on the ability of transformers to in-context learn a simple to study, yet general family of compositional functions: multi-layer perceptrons (MLPs). In this setting, we find that the success of CoT can be attributed to breaking down in-context learning of a compositional function into two distinct phases: focusing on and filtering data related to each step of the composition and in-context learning the single-step composition function. Through both experimental and theoretical evidence, we demonstrate how CoT significantly reduces the sample complexity of in-context learning (ICL) and facilitates the learning of complex functions that non-CoT methods struggle with. Furthermore, we illustrate how transformers can transition from vanilla in-context learning to mastering a compositional function with CoT by simply incorporating additional layers that perform the necessary data-filtering for CoT via the attention mechanism. In addition to these test-time benefits, we show CoT helps accelerate pretraining by learning shortcuts to represent complex functions and filtering plays an important role in this process. These findings collectively provide insights into the mechanics of CoT, inviting further investigation of its role in complex reasoning tasks.
Authors: Allan Raventós, Mansheej Paul, Feng Chen, Surya Ganguli
Pretrained transformers exhibit the remarkable ability of in-context learning (ICL): they can learn tasks from just a few examples provided in the prompt without updating any weights. This raises a foundational question: can ICL solve fundamentally $\textit{new}$ tasks that are very different from those seen during pretraining? To probe this question, we examine ICL's performance on linear regression while varying the diversity of tasks in the pretraining dataset. We empirically demonstrate a $\textit{task diversity threshold}$ for the emergence of ICL. Below this threshold, the pretrained transformer cannot solve unseen regression tasks, instead behaving like a Bayesian estimator with the $\textit{non-diverse pretraining task distribution}$ as the prior. Beyond this threshold, the transformer significantly outperforms this estimator; its behavior aligns with that of ridge regression, corresponding to a Gaussian prior over $\textit{all tasks}$, including those not seen during pretraining. Thus, when pretrained on data with task diversity greater than the threshold, transformers $\textit{can}$ optimally solve fundamentally new tasks in-context. Importantly, this capability hinges on it deviating from the Bayes optimal estimator with the pretraining distribution as the prior. This study also explores the effect of regularization, model capacity and task structure and underscores, in a concrete example, the critical role of task diversity, alongside data and model scale, in the emergence of ICL. Code is available at https://github.com/mansheej/icl-task-diversity.
Authors: Pierre Fernandez, Antoine Chaffin, Karim Tit, Vivien Chappelier, Teddy Furon
The task of discerning between generated and natural texts is increasingly challenging. In this context, watermarking emerges as a promising technique for ascribing generated text to a specific model. It alters the sampling generation process so as to leave an invisible trace in the generated output, facilitating later detection. This research consolidates watermarks for large language models based on three theoretical and empirical considerations. First, we introduce new statistical tests that offer robust theoretical guarantees which remain valid even at low false-positive rates (less than 10$^{\text{-6}}$). Second, we compare the effectiveness of watermarks using classical benchmarks in the field of natural language processing, gaining insights into their real-world applicability. Third, we develop advanced detection schemes for scenarios where access to the LLM is available, as well as multi-bit watermarking.
Authors: Serge Gladkoff, Gleb Erofeev, Irina Sorokina, Lifeng Han, Goran Nenadic
Translation Quality Evaluation (TQE) is an essential step of the modern translation production process. TQE is critical in assessing both machine translation (MT) and human translation (HT) quality without reference translations. The ability to evaluate or even simply estimate the quality of translation automatically may open significant efficiency gains through process optimisation. This work examines whether the state-of-the-art large language models (LLMs) can be used for this purpose. We take OpenAI models as the best state-of-the-art technology and approach TQE as a binary classification task. On eight language pairs including English to Italian, German, French, Japanese, Dutch, Portuguese, Turkish, and Chinese, our experimental results show that fine-tuned gpt3.5 can demonstrate good performance on translation quality prediction tasks, i.e. whether the translation needs to be edited. Another finding is that simply increasing the sizes of LLMs does not lead to apparent better performances on this task by comparing the performance of three different versions of OpenAI models: curie, davinci, and gpt3.5 with 13B, 175B, and 175B parameters, respectively.
Authors: Minsoo Kim, Sihwa Lee, Janghwan Lee, Sukjin Hong, Du-Seong Chang, Wonyong Sung, Jungwook Choi
Generative Language Models (GLMs) have shown impressive performance in tasks such as text generation, understanding, and reasoning. However, the large model size poses challenges for practical deployment. To solve this problem, Quantization-Aware Training (QAT) has become increasingly popular. However, current QAT methods for generative models have resulted in a noticeable loss of accuracy. To counteract this issue, we propose a novel knowledge distillation method specifically designed for GLMs. Our method, called token-scaled logit distillation, prevents overfitting and provides superior learning from the teacher model and ground truth. This research marks the first evaluation of ternary weight quantization-aware training of large-scale GLMs with less than 1.0 degradation in perplexity and achieves enhanced accuracy in tasks like common-sense QA and arithmetic reasoning as well as natural language understanding. Our code is available at https://github.com/aiha-lab/TSLD.
Authors: Omer Veysel Cagatan
We present ToddlerBERTa, a BabyBERTa-like language model, exploring its capabilities through five different models with varied hyperparameters. Evaluating on BLiMP, SuperGLUE, MSGS, and a Supplement benchmark from the BabyLM challenge, we find that smaller models can excel in specific tasks, while larger models perform well with substantial data. Despite training on a smaller dataset, ToddlerBERTa demonstrates commendable performance, rivalling the state-of-the-art RoBERTa-base. The model showcases robust language understanding, even with single-sentence pretraining, and competes with baselines that leverage broader contextual information. Our work provides insights into hyperparameter choices, and data utilization, contributing to the advancement of language models.
Authors: Hugh Zhang, David C. Parkes
Large language models have astounded the world with fascinating new capabilities. However, they currently lack the ability to teach themselves new skills, relying instead on large amounts of human-generated training data. We introduce SECToR (Self-Education via Chain-of-Thought Reasoning), a proof-of-concept demonstration that language models can teach themselves new skills using chain-of-thought reasoning. During the self-learning loop, SECToR asks models to solve addition problems using chain-of-thought reasoning before training the next version of the model to solve those same problems directly without using such reasoning. This process often results in an improved model which can, when again augmented with chain-of-thought reasoning, solve even harder problems than the original model, allowing the self-learning loop to continue. Language models trained via SECToR autonomously learn to add up to the longest-length-digit numbers without access to any ground truth examples beyond an initial supervised fine-tuning phase consisting only of numbers with 6 or fewer digits. Our central hypothesis is that chain-of-thought reasoning can act as a policy improvement operator, similarly to how Monte-Carlo Tree Search is used in AlphaZero (Silver et al., 2017). We hope that this research can lead to new directions in which language models can learn to teach themselves without the need for human demonstrations.
Authors: Huayang Li, Siheng Li, Deng Cai, Longyue Wang, Lemao Liu, Taro Watanabe, Yujiu Yang, Shuming Shi
Large language models with instruction-following abilities have revolutionized the field of artificial intelligence. These models show exceptional generalizability to tackle various real-world tasks through their natural language interfaces. However, their performance heavily relies on high-quality exemplar data, which is often difficult to obtain. This challenge is further exacerbated when it comes to multimodal instruction following. We introduce TextBind, an almost annotation-free framework for empowering larger language models with the multi-turn interleaved multimodal instruction-following capabilities. Our approach requires only image-caption pairs and generates multi-turn multimodal instruction-response conversations from a language model. To accommodate interleaved image-text inputs and outputs, we devise MIM, a language model-centric architecture that seamlessly integrates image encoder and decoder models. We release our dataset, model, and demo to foster future research in the area of multimodal instruction following.
Authors: Shiyi Zhu, Jing Ye, Wei Jiang, Qi Zhang, Yifan Wu, Jianguo Li
Self-attention and position embedding are two key modules in Transformer based LLMs. The potential relationship among them are far from well studied, especially for context window extending. In this paper, we introduce collinear constrained relationship to fuse RoPE and self-attention, and name it as Collinear Constrained Attention (CoCA). We've analyzed the computational and spatial complexity of CoCA and have determined that it adds only minimal additional overhead compared to the original Transformer-based models. We provide an efficient implementation of CoCA, and make it drop-in replacement for any existing position embedding and attention modules in Transformer based models. Experiments show that CoCA performs extraordinary well on context window extending. For instance, a CoCA based GPT model trained with 512 context length can extend the context window up to 8K without perplexity diverging. This indicates more than 16x context window extending without any fine-tuning. Our code is released here: https://github.com/codefuse-ai/Collinear-Constrained-Attention
Authors: Jon Saad-Falcon, Joe Barrow, Alexa Siu, Ani Nenkova, David Seunghyun Yoon, Ryan A. Rossi, Franck Dernoncourt
Large Language Models (LLMs) have issues with document question answering (QA) in situations where the document is unable to fit in the small context length of an LLM. To overcome this issue, most existing works focus on retrieving the relevant context from the document, representing them as plain text. However, documents such as PDFs, web pages, and presentations are naturally structured with different pages, tables, sections, and so on. Representing such structured documents as plain text is incongruous with the user's mental model of these documents with rich structure. When a system has to query the document for context, this incongruity is brought to the fore, and seemingly trivial questions can trip up the QA system. To bridge this fundamental gap in handling structured documents, we propose an approach called PDFTriage that enables models to retrieve the context based on either structure or content. Our experiments demonstrate the effectiveness of the proposed PDFTriage-augmented models across several classes of questions where existing retrieval-augmented LLMs fail. To facilitate further research on this fundamental problem, we release our benchmark dataset consisting of 900+ human-generated questions over 80 structured documents from 10 different categories of question types for document QA. Our code and datasets will be released soon on Github.
Authors: Xianming Li, Jing Li
High-quality text embedding is pivotal in improving semantic textual similarity (STS) tasks, which are crucial components in Large Language Model (LLM) applications. However, a common challenge existing text embedding models face is the problem of vanishing gradients, primarily due to their reliance on the cosine function in the optimization objective, which has saturation zones. To address this issue, this paper proposes a novel angle-optimized text embedding model called AnglE. The core idea of AnglE is to introduce angle optimization in a complex space. This novel approach effectively mitigates the adverse effects of the saturation zone in the cosine function, which can impede gradient and hinder optimization processes. To set up a comprehensive STS evaluation, we experimented on existing short-text STS datasets and a newly collected long-text STS dataset from GitHub Issues. Furthermore, we examine domain-specific STS scenarios with limited labeled data and explore how AnglE works with LLM-annotated data. Extensive experiments were conducted on various tasks including short-text STS, long-text STS, and domain-specific STS tasks. The results show that AnglE outperforms the state-of-the-art (SOTA) STS models that ignore the cosine saturation zone. These findings demonstrate the ability of AnglE to generate high-quality text embeddings and the usefulness of angle optimization in STS.
Authors: Yuzhang Shang, Zhihang Yuan, Qiang Wu, Zhen Dong
This paper explores network binarization, a radical form of quantization, compressing model weights to a single bit, specifically for Large Language Models (LLMs) compression. Due to previous binarization methods collapsing LLMs, we propose a novel approach, Partially-Binarized LLM (PB-LLM), which can achieve extreme low-bit quantization while maintaining the linguistic reasoning capacity of quantized LLMs. Specifically, our exploration first uncovers the ineffectiveness of naive applications of existing binarization algorithms and highlights the imperative role of salient weights in achieving low-bit quantization. Thus, PB-LLM filters a small ratio of salient weights during binarization, allocating them to higher-bit storage, i.e., partially-binarization. PB-LLM is extended to recover the capacities of quantized LMMs, by analyzing from the perspective of post-training quantization (PTQ) and quantization-aware training (QAT). Under PTQ, combining the concepts from GPTQ, we reconstruct the binarized weight matrix guided by the Hessian matrix and successfully recover the reasoning capacity of PB-LLM in low-bit. Under QAT, we freeze the salient weights during training, explore the derivation of optimal scaling factors crucial for minimizing the quantization error, and propose a scaling mechanism based on this derived scaling strategy for residual binarized weights. Those explorations and the developed methodologies significantly contribute to rejuvenating the performance of low-bit quantized LLMs and present substantial advancements in the field of network binarization for LLMs.The code is available at https://github.com/hahnyuan/BinaryLLM.
Authors: Jonathan Light, Min Cai, Sheng Shen, Ziniu Hu
In this paper, we explore the potential of Large Language Models (LLMs) Agents in playing the strategic social deduction game, Resistance Avalon. Players in Avalon are challenged not only to make informed decisions based on dynamically evolving game phases, but also to engage in discussions where they must deceive, deduce, and negotiate with other players. These characteristics make Avalon a compelling test-bed to study the decision-making and language-processing capabilities of LLM Agents. To facilitate research in this line, we introduce AvalonBench - a comprehensive game environment tailored for evaluating multi-agent LLM Agents. This benchmark incorporates: (1) a game environment for Avalon, (2) rule-based bots as baseline opponents, and (3) ReAct-style LLM agents with tailored prompts for each role. Notably, our evaluations based on AvalonBench highlight a clear capability gap. For instance, models like ChatGPT playing good-role got a win rate of 22.2% against rule-based bots playing evil, while good-role bot achieves 38.2% win rate in the same setting. We envision AvalonBench could be a good test-bed for developing more advanced LLMs (with self-playing) and agent frameworks that can effectively model the layered complexities of such game environments.
Authors: Lorenzo Canale, Alberto Messina
The Italian Digital Media Observatory (IDMO) project, part of a European initiative, focuses on countering disinformation and fake news. This report outlines contributions from Rai-CRITS to the project, including: (i) the creation of novel datasets for testing technologies (ii) development of an automatic model for categorizing Pagella Politica verdicts to facilitate broader analysis (iii) creation of an automatic model for recognizing textual entailment with exceptional accuracy on the FEVER dataset (iv) assessment using GPT-4 to identify textual entailmen (v) a game to raise awareness about fake news at national events.
Authors: Herbie Bradley, Andrew Dai, Hannah Teufel, Jenny Zhang, Koen Oostermeijer, Marco Bellagente, Jeff Clune, Kenneth Stanley, Grégory Schott, Joel Lehman
In many text-generation problems, users may prefer not only a single response, but a diverse range of high-quality outputs from which to choose. Quality-diversity (QD) search algorithms aim at such outcomes, by continually improving and diversifying a population of candidates. However, the applicability of QD to qualitative domains, like creative writing, has been limited by the difficulty of algorithmically specifying measures of quality and diversity. Interestingly, recent developments in language models (LMs) have enabled guiding search through AI feedback, wherein LMs are prompted in natural language to evaluate qualitative aspects of text. Leveraging this development, we introduce Quality-Diversity through AI Feedback (QDAIF), wherein an evolutionary algorithm applies LMs to both generate variation and evaluate the quality and diversity of candidate text. When assessed on creative writing domains, QDAIF covers more of a specified search space with high-quality samples than do non-QD controls. Further, human evaluation of QDAIF-generated creative texts validates reasonable agreement between AI and human evaluation. Our results thus highlight the potential of AI feedback to guide open-ended search for creative and original solutions, providing a recipe that seemingly generalizes to many domains and modalities. In this way, QDAIF is a step towards AI systems that can independently search, diversify, evaluate, and improve, which are among the core skills underlying human society's capacity for innovation.
Authors: David Q. Sun, Artem Abzaliev, Hadas Kotek, Zidi Xiu, Christopher Klein, Jason D. Williams
Controversy is a reflection of our zeitgeist, and an important aspect to any discourse. The rise of large language models (LLMs) as conversational systems has increased public reliance on these systems for answers to their various questions. Consequently, it is crucial to systematically examine how these models respond to questions that pertaining to ongoing debates. However, few such datasets exist in providing human-annotated labels reflecting the contemporary discussions. To foster research in this area, we propose a novel construction of a controversial questions dataset, expanding upon the publicly released Quora Question Pairs Dataset. This dataset presents challenges concerning knowledge recency, safety, fairness, and bias. We evaluate different LLMs using a subset of this dataset, illuminating how they handle controversial issues and the stances they adopt. This research ultimately contributes to our understanding of LLMs' interaction with controversial issues, paving the way for improvements in their comprehension and handling of complex societal debates.
Authors: Sajad Mousavi, Ricardo Luna Gutiérrez, Desik Rengarajan, Vineet Gundecha, Ashwin Ramesh Babu, Avisek Naug, Antonio Guillen, Soumyendu Sarkar
We propose a self-correction mechanism for Large Language Models (LLMs) to mitigate issues such as toxicity and fact hallucination. This method involves refining model outputs through an ensemble of critics and the model's own feedback. Drawing inspiration from human behavior, we explore whether LLMs can emulate the self-correction process observed in humans who often engage in self-reflection and seek input from others to refine their understanding of complex topics. Our approach is model-agnostic and can be applied across various domains to enhance trustworthiness by addressing fairness, bias, and robustness concerns. We consistently observe performance improvements in LLMs for reducing toxicity and correcting factual errors.
Authors: Md Tahmid Rahman Laskar, Xue-Yong Fu, Cheng Chen, Shashi Bhushan TN
This paper studies how to effectively build meeting summarization systems for real-world usage using large language models (LLMs). For this purpose, we conduct an extensive evaluation and comparison of various closed-source and open-source LLMs, namely, GPT-4, GPT- 3.5, PaLM-2, and LLaMA-2. Our findings reveal that most closed-source LLMs are generally better in terms of performance. However, much smaller open-source models like LLaMA- 2 (7B and 13B) could still achieve performance comparable to the large closed-source models even in zero-shot scenarios. Considering the privacy concerns of closed-source models for only being accessible via API, alongside the high cost associated with using fine-tuned versions of the closed-source models, the opensource models that can achieve competitive performance are more advantageous for industrial use. Balancing performance with associated costs and privacy concerns, the LLaMA-2-7B model looks more promising for industrial usage. In sum, this paper offers practical insights on using LLMs for real-world business meeting summarization, shedding light on the trade-offs between performance and cost.
Authors: Reza Khanmohammadi, Mohammad M. Ghassemi, Kyle Verdecchia, Ahmed I. Ghanem, Luo Bing, Indrin J. Chetty, Hassan Bagher-Ebadian, Farzan Siddiqui, Mohamed Elshaikh, Benjamin Movsas, Kundan Thind
Natural Language Processing (NLP) is a key technique for developing Medical Artificial Intelligence (AI) systems that leverage Electronic Health Record (EHR) data to build diagnostic and prognostic models. NLP enables the conversion of unstructured clinical text into structured data that can be fed into AI algorithms. The emergence of the transformer architecture and large language models (LLMs) has led to remarkable advances in NLP for various healthcare tasks, such as entity recognition, relation extraction, sentence similarity, text summarization, and question answering. In this article, we review the major technical innovations that underpin modern NLP models and present state-of-the-art NLP applications that employ LLMs in radiation oncology research. However, these LLMs are prone to many errors such as hallucinations, biases, and ethical violations, which necessitate rigorous evaluation and validation before clinical deployment. As such, we propose a comprehensive framework for assessing the NLP models based on their purpose and clinical fit, technical performance, bias and trust, legal and ethical implications, and quality assurance, prior to implementation in clinical radiation oncology. Our article aims to provide guidance and insights for researchers and clinicians who are interested in developing and using NLP models in clinical radiation oncology.
Authors: Lucas Georges Gabriel Charpentier, David Samuel
This paper introduces a novel modification of the transformer architecture, tailored for the data-efficient pretraining of language models. This aspect is evaluated by participating in the BabyLM challenge, where our solution won both the strict and strict-small tracks. Our approach allows each transformer layer to select which outputs of previous layers to process. The empirical results verify the potential of this simple modification and show that not all layers are equally as important.
Authors: Julien Guité-Vinet, Alexandre Blondin Massé, Fatiha Sadat
In the last years, several variants of transformers have emerged. In this paper, we compare different transformer-based models for solving the reverse dictionary task and explore their use in the context of a serious game called The Dictionary Game.
Authors: Harika Abburi, Kalyani Roy, Michael Suesserman, Nirmala Pudota, Balaji Veeramani, Edward Bowen, Sanmitra Bhattacharya
Recent Large Language Models (LLMs) have demonstrated remarkable capabilities in generating text that closely resembles human writing across wide range of styles and genres. However, such capabilities are prone to potential abuse, such as fake news generation, spam email creation, and misuse in academic assignments. Hence, it is essential to build automated approaches capable of distinguishing between artificially generated text and human-authored text. In this paper, we propose a simple yet efficient solution to this problem by ensembling predictions from multiple constituent LLMs. Compared to previous state-of-the-art approaches, which are perplexity-based or uses ensembles with a number of LLMs, our condensed ensembling approach uses only two constituent LLMs to achieve comparable performance. Experiments conducted on four benchmark datasets for generative text classification show performance improvements in the range of 0.5 to 100\% compared to previous state-of-the-art approaches. We also study the influence that the training data from individual LLMs have on model performance. We found that substituting commercially-restrictive Generative Pre-trained Transformer (GPT) data with data generated from other open language models such as Falcon, Large Language Model Meta AI (LLaMA2), and Mosaic Pretrained Transformers (MPT) is a feasible alternative when developing generative text detectors. Furthermore, to demonstrate zero-shot generalization, we experimented with an English essays dataset, and results suggest that our ensembling approach can handle new data effectively.
Authors: Jiale Cheng, Xiao Liu, Kehan Zheng, Pei Ke, Hongning Wang, Yuxiao Dong, Jie Tang, Minlie Huang
Large language models (LLMs) have shown impressive success in various applications. However, these models are often not well aligned with human intents, which calls for additional treatments on them, that is, the alignment problem. To make LLMs better follow user instructions, existing alignment methods mostly focus on further training them. However, the extra training of LLMs are usually expensive in terms of GPU compute; worse still, LLMs of interest are oftentimes not accessible for user-demanded training, such as GPTs. In this work, we take a different perspective -- Black-Box Prompt Optimization (BPO) -- to perform alignments. The idea is to optimize user prompts to suit LLMs' input understanding, so as to best realize users' intents without updating LLMs' parameters. BPO is model-agnostic and the empirical results demonstrate that the BPO-aligned ChatGPT yields a 22% increase in the win rate against its original version, and 10% for GPT-4. Importantly, the BPO-aligned LLMs can outperform the same models aligned by PPO and DPO, and it also brings additional performance gains when combining BPO with PPO or DPO. Code and datasets are released at https://github.com/thu-coai/BPO.