This is an internal, incomplete preview of a proposed change to the ACL Anthology.
For efficiency reasons, we generate only three BibTeX files per volume, and the preview may be incomplete in other ways, or contain mistakes.
Do not treat this content as an official publication.
With the increasing capabilities of large language models (LLMs), in-context learning (ICL) has emerged as a new paradigm for natural language processing (NLP), where LLMs make predictions based on contexts augmented with a few examples. It has been a significant trend to explore ICL to evaluate and extrapolate the ability of LLMs. In this paper, we aim to survey and summarize the progress and challenges of ICL. We first present a formal definition of ICL and clarify its correlation to related studies. Then, we organize and discuss advanced techniques, including training strategies, prompt designing strategies, and related analysis. Additionally, we explore various ICL application scenarios, such as data engineering and knowledge updating. Finally, we address the challenges of ICL and suggest potential directions for further research. We hope that our work can encourage more research on uncovering how ICL works and improving ICL.
Large language models (LLMs) have demonstrated impressive capabilities, but still suffer from inconsistency issues (e.g. LLMs can react differently to disturbances like rephrasing or inconsequential order change). In addition to these inconsistencies, we also observe that LLMs, while capable of solving hard problems, can paradoxically fail at easier ones. To evaluate this hard-to-easy inconsistency, we develop the ConsisEval benchmark, where each entry comprises a pair of questions with a strict order of difficulty. Furthermore, we introduce the concept of consistency score to quantitatively measure this inconsistency and analyze the potential for improvement in consistency by relative consistency score. Based on comprehensive experiments across a variety of existing models, we find: (1) GPT-4 achieves the highest consistency score of 92.2% but is still inconsistent to specific questions due to distraction by redundant information, misinterpretation of questions, etc.; (2) models with stronger capabilities typically exhibit higher consistency, but exceptions also exist; (3) hard data enhances consistency for both fine-tuning and in-context learning. Our data and code will be publicly available on GitHub.
Understanding the deep semantics of images is essential in the era dominated by social media. However, current research works primarily on the superficial description of images, revealing a notable deficiency in the systematic investigation of the inherent deep semantics. In this work, we introduce DEEPEVAL, a comprehensive benchmark to assess Large Multimodal Models’ (LMMs) capacities of visual deep semantics. DEEPEVAL includes human-annotated dataset and three progressive subtasks: fine-grained description selection, in-depth title matching, and deep semantics understanding. Utilizing DEEPEVAL, we evaluate 9 open-source LMMs and GPT-4V(ision). Our evaluation demonstrates a substantial gap between the deep semantic comprehension capabilities of existing LMMs and humans. For example, GPT-4V is 30% behind humans in understanding deep semantics, even though it achieves human-comparable performance in image description. Further analysis reveals that LMM performance on DEEPEVAL varies according to the specific facets of deep semantics explored, indicating the fundamental challenges remaining in developing LMMs.
With promising yet saturated results in high-resource settings, low-resource datasets have gradually become crucial benchmarks (e.g., BigBench Hard, superGLUE) for evaluating the learning ability of advanced neural networks. In this work, we find that there exists a set of “hard examples” in low-resource settings that challenge neural networks but are not well evaluated, which causes over-estimated performance. We first give a theoretical analysis on which factors bring the difficulty of low-resource learning. It then motivates us to propose a challenging benchmark Achilles-Bench to better evaluate the learning ability, which covers 11 datasets, including 8 natural language process (NLP) datasets and 3 computer vision (CV) datasets. Experiments on a wide range of models show that neural networks, even pre-trained language models, have sharp performance drops on our benchmark, demonstrating the effectiveness of evaluating the weaknesses of neural networks. On NLP tasks, we surprisingly find that despite better results on traditional low-resource benchmarks, pre-trained networks, does not show performance improvements on our benchmarks. there is still a large robustness gap between existing models and human-level performance, highlighting the need for robust low-resource learning models.
To mitigate the high inference latency stemming from autoregressive decoding in Large Language Models (LLMs), Speculative Decoding has emerged as a novel decoding paradigm for LLM inference. In each decoding step, this method first drafts several future tokens efficiently and then verifies them in parallel. Unlike autoregressive decoding, Speculative Decoding facilitates the simultaneous decoding of multiple tokens per step, thereby accelerating inference. This paper presents a comprehensive overview and analysis of this promising decoding paradigm. We begin by providing a formal definition and formulation of Speculative Decoding. Then, we organize in-depth discussions on its key facets, such as drafter selection and verification strategies. Furthermore, we present a comparative analysis of leading methods under third-party testing environments. We aim for this work to serve as a catalyst for further research on Speculative Decoding, ultimately contributing to more efficient LLM inference.
Large Language Models (LLMs) have gained increasing attention for their remarkable capacity, alongside concerns about safety arising from their potential to produce harmful content. Red teaming aims to find prompts that could elicit harmful responses from LLMs, and is essential to discover and mitigate safety risks before real-world deployment. However, manual red teaming is both time-consuming and expensive, rendering it unscalable. In this paper, we propose RTPE, a scalable evolution framework to evolve red teaming prompts across both breadth and depth dimensions, facilitating the automatic generation of numerous high-quality and diverse red teaming prompts. Specifically, in-breadth evolving employs a novel enhanced in-context learning method to create a multitude of quality prompts, whereas in-depth evolving applies customized transformation operations to enhance both content and form of prompts, thereby increasing diversity. Extensive experiments demonstrate that RTPE surpasses existing representative automatic red teaming methods on both attack success rate and diversity. In addition, based on 4,800 red teaming prompts created by RTPE, we further provide a systematic analysis of 8 representative LLMs across 8 sensitive topics.
Large language models (LLMs) have shown promising efficacy across various tasks, becoming powerful tools in numerous aspects of human life. However, Transformer-based LLMs suffer a performance degradation when modeling long-term contexts due to they discard some information to reduce computational overhead. In this work, we propose a simple yet effective method to enable LLMs to take a deep breath, encouraging them to summarize information contained within discrete text chunks. Specifically, we segment the text into multiple chunks and insert special token <SR> at the end of each chunk. We then modify the attention mask to integrate the chunk’s information into the corresponding <SR> token. This facilitates LLMs to interpret information not only from historical individual tokens but also from the <SR> token, aggregating the chunk’s semantic information. Experiments on language modeling and out-of-domain downstream tasks validate the superiority of our approach.
The safety of Large Language Models (LLMs) has gained increasing attention in recent years, but there still lacks a comprehensive approach for detecting safety issues within LLMs’ responses in an aligned, customizable and explainable manner. In this paper, we propose ShieldLM, an LLM-based safety detector, which aligns with common safety standards, supports customizable detection rules, and provides explanations for its decisions. To train ShieldLM, we compile a large bilingual dataset comprising 14,387 query-response pairs, annotating the safety of responses based on various safety standards. Through extensive experiments, we demonstrate that ShieldLM surpasses strong baselines across four test sets, showcasing remarkable customizability and explainability. Besides performing well on standard detection datasets, ShieldLM has also been shown to be effective as a safety evaluator for advanced LLMs. ShieldLM is released at https://github.com/thu-coai/ShieldLM to support accurate and explainable safety detection under various safety standards.
In the era of large language models, Mixture-of-Experts (MoE) is a promising architecture for managing computational costs when scaling up model parameters. However, conventional MoE architectures like GShard, which activate the top-K out of N experts, face challenges in ensuring expert specialization, i.e. each expert acquires non-overlapping and focused knowledge. In response, we propose the DeepSeekMoE architecture towards ultimate expert specialization. It involves two principal strategies: (1) finely segmenting the experts into mN ones and activating mK from them, allowing for a more flexible combination of activated experts; (2) isolating Ks experts as shared ones, aiming at capturing common knowledge and mitigating redundancy in routed experts. Starting from a modest scale with 2B parameters, we demonstrate that DeepSeekMoE 2B achieves comparable performance with GShard 2.9B, which has 1.5 × expert parameters and computation. In addition, DeepSeekMoE 2B nearly approaches the performance of its dense counterpart with the same number of total parameters, which sets the upper bound of MoE models. Subsequently, we scale up DeepSeekMoE to 16B parameters and show that it achieves comparable performance with DeepSeek 7B and LLaMA2 7B, with only about 40% of computations.
In this paper, we present an innovative process-oriented math process reward model called Math-shepherd, which assigns a reward score to each step of math problem solutions. The training of Math-shepherd is achieved using automatically constructed process-wise supervision data, breaking the bottleneck of heavy reliance on manual annotation in existing work. We explore the effectiveness of Math-shepherd in two scenarios: 1) Verification: Math-shepherd is utilized for reranking multiple outputs generated by Large Language Models (LLMs); 2) Reinforcement Learning (RL): Math-shepherd is employed to reinforce LLMs.With Math-shepherd, a series of open-source LLMs demonstrates exceptional performance. For instance, process RL with Math-shepherd significantly enhances Mistral-7B (77.9%→84.1% on GSM8K and 28.6%→33.0% on MATH).The accuracy can be further improved to 89.1% and 43.5% on two benchmarks with verification of Math-shepherd.We believe that automatic process supervision holds significant potential for the future evolution of LLMs.
In this paper, we uncover a positional bias in the evaluation paradigm of adopting large language models (LLMs), e.g., GPT-4, as a referee to score and compare the quality of responses generated by candidate models. We find that the quality ranking of candidate responses can be easily hacked by simply altering their order of appearance in the context. This manipulation allows us to skew the evaluation result, making one model appear considerably superior to the other, e.g., Vicuna-13B could beat ChatGPT on 66 over 80 tested queries with ChatGPT as an evaluator. We propose a simple yet effective calibration framework to address our discovered positional bias.To evaluate the effectiveness of our framework, we manually annotate the “win/tie/lose” outcomes of responses from ChatGPT and Vicuna-13B in the Vicuna Benchmark’s question prompt. Extensive experiments demonstrate that our approach successfully alleviates evaluation bias, resulting in closer alignment with human judgments.
In the face of the rapidly growing spread of false and misleading information in the real world, manual evidence-based fact-checking efforts become increasingly challenging and time-consuming. In order to tackle this issue, we propose FaGANet, an automated and accurate fact-checking model that leverages the power of sentence-level attention and graph attention network to enhance performance. This model adeptly integrates encoder-only models with graph attention network, effectively fusing claims and evidence information for accurate identification of even well-disguised data. Experiment results showcase the significant improvement in accuracy achieved by our FaGANet model, as well as its state-of-the-art performance in the evidence-based fact-checking task. We release our code and data in https://github.com/WeiyaoLuo/FaGANet.
Video multimodal fusion aims to integrate multimodal signals in videos, such as visual, audio and text, to make a complementary prediction with multiple modalities contents. However, unlike other image-text multimodal tasks, video has longer multimodal sequences with more redundancy and noise in both visual and audio modalities. Prior denoising methods like forget gate are coarse in the granularity of noise filtering. They often suppress the redundant and noisy information at the risk of losing critical information. Therefore, we propose a denoising bottleneck fusion (DBF) model for fine-grained video multimodal fusion. On the one hand, we employ a bottleneck mechanism to filter out noise and redundancy with a restrained receptive field. On the other hand, we use a mutual information maximization module to regulate the filter-out module to preserve key information within different modalities. Our DBF model achieves significant improvement over current state-of-the-art baselines on multiple benchmarks covering multimodal sentiment analysis and multimodal summarization tasks. It proves that our model can effectively capture salient features from noisy and redundant video, audio, and text inputs. The code for this paper will be publicly available at https://github.com/WSXRHFG/DBF
Large pretrained language models have shown surprising in-context learning (ICL) ability. With a few demonstration input-label pairs, they can predict the label for an unseen input without parameter updates. Despite the great success in performance, its working mechanism still remains an open question. In this paper, we explain language models as meta-optimizers and understand in-context learning as implicit finetuning. Theoretically, we figure out that Transformer attention has a dual form of gradient descent. On top of it, we understand ICL as follows: GPT first produces meta-gradients according to the demonstration examples, and then these meta-gradients are applied to the original GPT to build an ICL model. We comprehensively compare the behaviors of in-context learning and explicit finetuning on real tasks to provide empirical evidence that supports our understanding. Experimental results show that in-context learning behaves similarly to explicit finetuning from multiple perspectives. Inspired by the dual form between Transformer attention and gradient descent, we design a momentum-based attention by analogy with gradient descent with momentum. The improved performance over vanilla attention further supports our understanding from another perspective, and more importantly, shows the potential to utilize our understanding for future model design. The code is available at https://aka.ms/icl.
Continual relation extraction (CRE) models aim at handling emerging new relations while avoiding catastrophically forgetting old ones in the streaming data. Though improvements have been shown by previous CRE studies, most of them only adopt a vanilla strategy when models first learn representations of new relations. In this work, we point out that there exist two typical biases after training of this vanilla strategy: classifier bias and representation bias, which causes the previous knowledge that the model learned to be shaded. To alleviate those biases, we propose a simple yet effective classifier decomposition framework that splits the last FFN layer into separated previous and current classifiers, so as to maintain previous knowledge and encourage the model to learn more robust representations at this training stage. Experimental results on two standard benchmarks show that our proposed framework consistently outperforms the state-of-the-art CRE models, which indicates that the importance of the first training stage to CRE models may be underestimated. Our code will be released upon acceptance.
Entity linking models have achieved significant success via utilizing pretrained language models to capture semantic features. However, the NIL prediction problem, which aims to identify mentions without a corresponding entity in the knowledge base, has received insufficient attention. We categorize mentions linking to NIL into Missing Entity and Non-Entity Phrase, and propose an entity linking dataset NEL that focuses on the NIL prediction problem.NEL takes ambiguous entities as seeds, collects relevant mention context in the Wikipedia corpus, and ensures the presence of mentions linking to NIL by human annotation and entity masking. We conduct a series of experiments with the widely used bi-encoder and cross-encoder entity linking models, results show that both types of NIL mentions in training data have a significant influence on the accuracy of NIL prediction. Our code and dataset can be accessed at https://github.com/solitaryzero/NIL_EL.
Abstract Meaning Representation (AMR) parsing aims to extract an abstract semantic graph from a given sentence. The sequence-to-sequence approaches, which linearize the semantic graph into a sequence of nodes and edges and generate the linearized graph directly, have achieved good performance. However, we observed that these approaches suffer from structure loss accumulation during the decoding process, leading to a much lower F1-score for nodes and edges decoded later compared to those decoded earlier. To address this issue, we propose a novel Reverse Graph Linearization (RGL) enhanced framework. RGL defines both default and reverse linearization orders of an AMR graph, where most structures at the back part of the default order appear at the front part of the reversed order and vice versa. RGL incorporates the reversed linearization to the original AMR parser through a two-pass self-distillation mechanism, which guides the model when generating the default linearizations. Our analysis shows that our proposed method significantly mitigates the problem of structure loss accumulation, outperforming the previously best AMR parsing model by 0.8 and 0.5 Smatch scores on the AMR 2.0 and AMR 3.0 dataset, respectively. The code are available at https://github.com/pkunlp-icler/AMR_reverse_graph_linearization.
Recently, Large Language Models (LLMs) have been serving as general-purpose interfaces, posing a significant demand for comprehensive visual knowledge. However, it remains unclear how well current LLMs and their visually augmented counterparts (VaLMs) can master visual commonsense knowledge. To investigate this, we propose ImageNetVC, a human-annotated dataset specifically designed for zero- and few-shot visual commonsense evaluation across 1,000 ImageNet categories. Utilizing ImageNetVC, we benchmark the fundamental visual commonsense knowledge of both unimodal LLMs and VaLMs. Furthermore, we analyze the factors affecting the visual commonsense knowledge of large-scale models, providing insights into the development of language models enriched with visual commonsense knowledge. Our code and dataset are available at https://github.com/hemingkx/ImageNetVC.
We propose Speculative Decoding (SpecDec), for the first time ever, to formally study exploiting the idea of speculative execution to accelerate autoregressive (AR) decoding. Speculative Decoding has two innovations: Spec-Drafter – an independent model specially optimized for efficient and accurate drafting – and Spec-Verification – a reliable method for verifying the drafted tokens efficiently in the decoding paradigm. Experimental results on various seq2seq tasks including machine translation and abstractive summarization show our approach can achieve around 5x speedup for the popular Transformer architectures with comparable generation quality to beam search decoding, refreshing the impression that the draft-then-verify paradigm introduces only 1.4x~2x speedup. In addition to the remarkable speedup, we also demonstrate 3 additional advantages of SpecDec, revealing its practical value for accelerating generative models in real-world applications. Our models and codes are available at https://github.com/hemingkx/SpecDec.
Pretrained language models have achieved remarkable success in natural language understanding. However, fine-tuning pretrained models on limited training data tends to overfit and thus diminish performance. This paper presents Bi-Drop, a fine-tuning strategy that selectively updates model parameters using gradients from various sub-nets dynamically generated by dropout. The sub-net estimation of Bi-Drop is performed in an in-batch manner, so it overcomes the problem of hysteresis in sub-net updating, which is possessed by previous methods that perform asynchronous sub-net estimation. Also, Bi-Drop needs only one mini-batch to estimate the sub-net so it achieves higher utility of training data. Experiments on the GLUE benchmark demonstrate that Bi-Drop consistently outperforms previous fine-tuning methods. Furthermore, empirical results also show that Bi-Drop exhibits excellent generalization ability and robustness for domain transfer, data imbalance, and low-resource scenarios.
Harvesting question-answer (QA) pairs from customer service chatlog in the wild is an efficient way to enrich the knowledge base for customer service chatbots in the cold start or continuous integration scenarios. Prior work attempts to obtain 1-to-1 QA pairs from growing customer service chatlog, which fails to integrate the incomplete utterances from the dialog context for composite QA retrieval. In this paper, we propose N-to-N QA extraction task in which the derived questions and corresponding answers might be separated across different utterances. We introduce a suite of generative/discriminative tagging based methods with end-to-end and two-stage variants that perform well on 5 customer service datasets and for the first time setup a benchmark for N-to-N DialogQAE with utterance and session level evaluation metrics. With a deep dive into extracted QA pairs, we find that the relations between and inside the QA pairs can be indicators to analyze the dialogue structure, e.g. information seeking, clarification, barge-in and elaboration. We also show that the proposed models can adapt to different domains and languages, and reduce the labor cost of knowledge accumulation in the real-world product dialogue platform.
Large Language Models (LLMs) have recently gained the In-Context Learning (ICL) ability with the models scaling up, allowing them to quickly adapt to downstream tasks with only a few demonstration examples prepended in the input sequence. Nonetheless, the current practice of ICL treats all demonstration examples equally, which still warrants improvement, as the quality of examples is usually uneven. In this paper, we investigate how to determine approximately optimal weights for demonstration examples and how to apply them during ICL. To assess the quality of weights in the absence of additional validation data, we design a masked self-prediction (MSP) score that exhibits a strong correlation with the final ICL performance. To expedite the weight-searching process, we discretize the continuous weight space and adopt beam search. With approximately optimal weights obtained, we further propose two strategies to apply them to demonstrations at different model positions. Experimental results on 8 text classification tasks show that our approach outperforms conventional ICL by a large margin. Our code are publicly available at https:github.com/Zhe-Young/WICL.
Continual learning (CL) aims to constantly learn new knowledge over time while avoiding catastrophic forgetting on old tasks. We focus on continual text classification under the class-incremental setting. Recent CL studies have identified the severe performance decrease on analogous classes as a key factor for catastrophic forgetting. In this paper, through an in-depth exploration of the representation learning process in CL, we discover that the compression effect of the information bottleneck leads to confusion on analogous classes. To enable the model learn more sufficient representations, we propose a novel replay-based continual text classification method, InfoCL. Our approach utilizes fast-slow and current-past contrastive learning to perform mutual information maximization and better recover the previously learned representations. In addition, InfoCL incorporates an adversarial memory augmentation strategy to alleviate the overfitting problem of replay. Experimental results demonstrate that InfoCL effectively mitigates forgetting and achieves state-of-the-art performance on three text classification tasks.
Few-Shot Sequence Labeling (FSSL) is a canonical paradigm for the tagging models, e.g., named entity recognition and slot filling, to generalize on an emerging, resource-scarce domain. Recently, the metric-based meta-learning framework has been recognized as a promising approach for FSSL. However, most prior works assign a label to each token based on the token-level similarities, which ignores the integrality of named entities or slots. To this end, in this paper, we propose ESD, an Enhanced Span-based Decomposition method for FSSL. ESD formulates FSSL as a span-level matching problem between test query and supporting instances. Specifically, ESD decomposes the span matching problem into a series of span-level procedures, mainly including enhanced span representation, class prototype aggregation and span conflicts resolution. Extensive experiments show that ESD achieves the new state-of-the-art results on two popular FSSL benchmarks, FewNERD and SNIPS, and is proven to be more robust in the noisy and nested tagging scenarios.
Most previous studies aim at extracting events from a single sentence, while document-level event extraction still remains under-explored. In this paper, we focus on extracting event arguments from an entire document, which mainly faces two critical problems: a) the long-distance dependency between trigger and arguments over sentences; b) the distracting context towards an event in the document. To address these issues, we propose a Two-Stream Abstract meaning Representation enhanced extraction model (TSAR). TSAR encodes the document from different perspectives by a two-stream encoding module, to utilize local and global information and lower the impact of distracting context. Besides, TSAR introduces an AMR-guided interaction module to capture both intra-sentential and inter-sentential features, based on the locally and globally constructed AMR semantic graphs. An auxiliary boundary loss is introduced to enhance the boundary information for text spans explicitly. Extensive experiments illustrate that TSAR outperforms previous state-of-the-art by a large margin, with 2.54 F1 and 5.13 F1 performance gain on the public RAMS and WikiEvents datasets respectively, showing the superiority in the cross-sentence arguments extraction. We release our code in https://github.com/PKUnlp-icler/TSAR.
It is a common practice for recent works in vision language cross-modal reasoning to adopt a binary or multi-choice classification formulation taking as input a set of source image(s) and textual query. In this work, we take a sober look at such an “unconditional” formulation in the sense that no prior knowledge is specified with respect to the source image(s). Inspired by the designs of both visual commonsense reasoning and natural language inference tasks, we propose a new task termed “Premise-based Multi-modal Reasoning” (PMR) where a textual premise is the background presumption on each source image. The PMR dataset contains 15,360 manually annotated samples which are created by a multi-phase crowd-sourcing process. With selected high-quality movie screenshots and human-curated premise templates from 6 pre-defined categories, we ask crowd-source workers to write one true hypothesis and three distractors (4 choices) given the premise and image through a cross-check procedure.
Large pretrained generative models like GPT-3 often suffer from hallucinating non-existent or incorrect content, which undermines their potential merits in real applications. Existing work usually attempts to detect these hallucinations based on a corresponding oracle reference at a sentence or document level. However ground-truth references may not be readily available for many free-form text generation applications, and sentence- or document-level detection may fail to provide the fine-grained signals that would prevent fallacious content in real time. As a first step to addressing these issues, we propose a novel token-level, reference-free hallucination detection task and an associated annotated dataset named HaDeS (HAllucination DEtection dataSet). To create this dataset, we first perturb a large number of text segments extracted from English language Wikipedia, and then verify these with crowd-sourced annotations. To mitigate label imbalance during annotation, we utilize an iterative model-in-loop strategy. We conduct comprehensive data analyses and create multiple baseline models.
The Mixture-of-Experts (MoE) technique can scale up the model size of Transformers with an affordable computational overhead. We point out that existing learning-to-route MoE methods suffer from the routing fluctuation issue, i.e., the target expert of the same input may change along with training, but only one expert will be activated for the input during inference. The routing fluctuation tends to harm sample efficiency because the same input updates different experts but only one is finally used. In this paper, we propose StableMoE with two training stages to address the routing fluctuation problem. In the first training stage, we learn a balanced and cohesive routing strategy and distill it into a lightweight router decoupled from the backbone model. In the second training stage, we utilize the distilled router to determine the token-to-expert assignment and freeze it for a stable routing strategy. We validate our method on language modeling and multilingual machine translation. The results show that StableMoE outperforms existing MoE methods in terms of both convergence speed and performance.
Artificial Intelligence (AI), along with the recent progress in biomedical language understanding, is gradually offering great promise for medical practice. With the development of biomedical language understanding benchmarks, AI applications are widely used in the medical field. However, most benchmarks are limited to English, which makes it challenging to replicate many of the successes in English for other languages. To facilitate research in this direction, we collect real-world biomedical data and present the first Chinese Biomedical Language Understanding Evaluation (CBLUE) benchmark: a collection of natural language understanding tasks including named entity recognition, information extraction, clinical diagnosis normalization, single-sentence/sentence-pair classification, and an associated online platform for model evaluation, comparison, and analysis. To establish evaluation on these tasks, we report empirical results with the current 11 pre-trained Chinese models, and experimental results show that state-of-the-art neural models perform by far worse than the human ceiling.
Large-scale pretrained language models are surprisingly good at recalling factual knowledge presented in the training corpus. In this paper, we present preliminary studies on how factual knowledge is stored in pretrained Transformers by introducing the concept of knowledge neurons. Specifically, we examine the fill-in-the-blank cloze task for BERT. Given a relational fact, we propose a knowledge attribution method to identify the neurons that express the fact. We find that the activation of such knowledge neurons is positively correlated to the expression of their corresponding facts. In our case studies, we attempt to leverage knowledge neurons to edit (such as update, and erase) specific factual knowledge without fine-tuning. Our results shed light on understanding the storage of knowledge within pretrained Transformers.
Abstract Meaning Representation (AMR) parsing aims to translate sentences to semantic representation with a hierarchical structure, and is recently empowered by pretrained sequence-to-sequence models. However, there exists a gap between their flat training objective (i.e., equally treats all output tokens) and the hierarchical AMR structure, which limits the model generalization. To bridge this gap, we propose a Hierarchical Curriculum Learning (HCL) framework with Structure-level (SC) and Instance-level Curricula (IC). SC switches progressively from core to detail AMR semantic elements while IC transits from structure-simple to -complex AMR instances during training. Through these two warming-up processes, HCL reduces the difficulty of learning complex structures, thus the flat model can better adapt to the AMR hierarchy. Extensive experiments on AMR2.0, AMR3.0, structure-complex and out-of-distribution situations verify the effectiveness of HCL.
Hierarchical text classification (HTC) is a challenging subtask of multi-label classification due to its complex label hierarchy.Recently, the pretrained language models (PLM)have been widely adopted in HTC through a fine-tuning paradigm. However, in this paradigm, there exists a huge gap between the classification tasks with sophisticated label hierarchy and the masked language model (MLM) pretraining tasks of PLMs and thus the potential of PLMs cannot be fully tapped.To bridge the gap, in this paper, we propose HPT, a Hierarchy-aware Prompt Tuning method to handle HTC from a multi-label MLM perspective.Specifically, we construct a dynamic virtual template and label words that take the form of soft prompts to fuse the label hierarchy knowledge and introduce a zero-bounded multi-label cross-entropy loss to harmonize the objectives of HTC and MLM.Extensive experiments show HPT achieves state-of-the-art performances on 3 popular HTC datasets and is adept at handling the imbalance and low resource situations. Our code is available at https://github.com/wzh9969/HPT.
Continual relation extraction (CRE) aims to continually learn new relations from a class-incremental data stream. CRE model usually suffers from catastrophic forgetting problem, i.e., the performance of old relations seriously degrades when the model learns new relations. Most previous work attributes catastrophic forgetting to the corruption of the learned representations as new relations come, with an implicit assumption that the CRE models have adequately learned the old relations. In this paper, through empirical studies we argue that this assumption may not hold, and an important reason for catastrophic forgetting is that the learned representations do not have good robustness against the appearance of analogous relations in the subsequent learning process. To address this issue, we encourage the model to learn more precise and robust representations through a simple yet effective adversarial class augmentation mechanism (ACA), which is easy to implement and model-agnostic.Experimental results show that ACA can consistently improve the performance of state-of-the-art CRE models on two popular benchmarks.
As Abstract Meaning Representation (AMR) implicitly involves compound semantic annotations, we hypothesize auxiliary tasks which are semantically or formally related can better enhance AMR parsing. We find that 1) Semantic role labeling (SRL) and dependency parsing (DP), would bring more performance gain than other tasks e.g. MT and summarization in the text-to-AMR transition even with much less data. 2) To make a better fit for AMR, data from auxiliary tasks should be properly “AMRized” to PseudoAMR before training. Knowledge from shallow level parsing tasks can be better transferred to AMR Parsing with structure transform. 3) Intermediate-task learning is a better paradigm to introduce auxiliary tasks to AMR parsing, compared to multitask learning. From an empirical perspective, we propose a principled method to involve auxiliary tasks to boost AMR parsing. Extensive experiments show that our method achieves new state-of-the-art performance on different benchmarks especially in topology-related scores. Code and models are released at https://github.com/PKUnlp-icler/ATP.
While interacting with chatbots, users may elicit multiple intents in a single dialogue utterance. Instead of training a dedicated multi-intent detection model, we propose DialogUSR, a dialogue utterance splitting and reformulation task that first splits multi-intent user query into several single-intent sub-queries and then recovers all the coreferred and omitted information in the sub-queries. DialogUSR can serve as a plug-in and domain-agnostic module that empowers the multi-intent detection for the deployed chatbots with minimal efforts. We collect a high-quality naturally occurring dataset that covers 23 domains with a multi-step crowd-souring procedure. To benchmark the proposed dataset, we propose multiple action-based generative models that involve end-to-end and two-stage training, and conduct in-depth analyses on the pros and cons of the proposed baselines.
Previous literature has proved that Pretrained Language Models (PLMs) can store factual knowledge. However, we find that facts stored in the PLMs are not always correct. It motivates us to explore a fundamental question: How do we calibrate factual knowledge in PLMs without re-training from scratch? In this work, we propose a simple and lightweight method CaliNet to achieve this goal. To be specific, we first detect whether PLMs can learn the right facts via a contrastive score between right and fake facts. If not, we then use a lightweight method to add and adapt new parameters to specific factual texts. Experiments on the knowledge probing task show the calibration effectiveness and efficiency. In addition, through closed-book question answering, we find that the calibrated PLM possesses knowledge generalization ability after finetuning.Beyond the calibration performance, we further investigate and visualize the knowledge calibration mechanism.
In this paper, we tackle the task of Definition Generation (DG) in Chinese, which aims at automatically generating a definition for a word. Most existing methods take the source word as an indecomposable semantic unit. However, in parataxis languages like Chinese, word meanings can be composed using the word formation process, where a word (“桃花”, peach-blossom) is formed by formation components (“桃”, peach; “花”, flower) using a formation rule (Modifier-Head). Inspired by this process, we propose to enhance DG with word formation features. We build a formation-informed dataset, and propose a model DeFT, which Decomposes words into formation features, dynamically Fuses different features through a gating mechanism, and generaTes word definitions. Experimental results show that our method is both effective and robust.
Conventional Knowledge Graph Completion (KGC) assumes that all test entities appear during training. However, in real-world scenarios, Knowledge Graphs (KG) evolve fast with out-of-knowledge-graph (OOKG) entities added frequently, and we need to efficiently represent these entities. Most existing Knowledge Graph Embedding (KGE) methods cannot represent OOKG entities without costly retraining on the whole KG. To enhance efficiency, we propose a simple and effective method that inductively represents OOKG entities by their optimal estimation under translational assumptions. Moreover, given pretrained embeddings of the in-knowledge-graph (IKG) entities, our method even needs no additional learning. Experimental results on two KGC tasks with OOKG entities show that our method outperforms the previous methods by a large margin with higher efficiency.
The widespread adoption of reference-based automatic evaluation metrics such as ROUGE has promoted the development of document summarization. In this paper, we consider a new protocol for designing reference-based metrics that require the endorsement of source document(s). Following protocol, we propose an anchored ROUGE metric fixing each summary particle on source document, which bases the computation on more solid ground. Empirical results on benchmark datasets validate that source document helps to induce a higher correlation with human judgments for ROUGE metric. Being self-explanatory and easy-to-implement, the protocol can naturally foster various effective designs of reference-based metrics besides the anchored ROUGE introduced here.
Many recent studies have shown that for models trained on datasets for natural language inference (NLI), it is possible to make correct predictions by merely looking at the hypothesis while completely ignoring the premise. In this work, we manage to derive adversarial examples in terms of the hypothesis-only bias and explore eligible ways to mitigate such bias. Specifically, we extract various phrases from the hypotheses (artificial patterns) in the training sets, and show that they have been strong indicators to the specific labels. We then figure out ‘hard’ and ‘easy’ instances from the original test sets whose labels are opposite to or consistent with those indications. We also set up baselines including both pretrained models (BERT, RoBerta, XLNet) and competitive non-pretrained models (InferSent, DAM, ESIM). Apart from the benchmark and baselines, we also investigate two debiasing approaches which exploit the artificial pattern modeling to mitigate such hypothesis-only bias: down-sampling and adversarial training. We believe those methods can be treated as competitive baselines in NLI debiasing tasks.
The prior work on natural language inference (NLI) debiasing mainly targets at one or few known biases while not necessarily making the models more robust. In this paper, we focus on the model-agnostic debiasing strategies and explore how to (or is it possible to) make the NLI models robust to multiple distinct adversarial attacks while keeping or even strengthening the models’ generalization power. We firstly benchmark prevailing neural NLI models including pretrained ones on various adversarial datasets. We then try to combat distinct known biases by modifying a mixture of experts (MoE) ensemble method and show that it’s nontrivial to mitigate multiple NLI biases at the same time, and that model-level ensemble method outperforms MoE ensemble method. We also perform data augmentation including text swap, word substitution and paraphrase and prove its efficiency in combating various (though not all) adversarial attacks at the same time. Finally, we investigate several methods to merge heterogeneous training data (1.35M) and perform model ensembling, which are straightforward but effective to strengthen NLI models.
Multi-document summarization (MDS) aims at producing a good-quality summary for several related documents. In this paper, we propose a spectral-based hypothesis, which states that the goodness of summary candidate is closely linked to its so-called spectral impact. Here spectral impact considers the perturbation to the dominant eigenvalue of affinity matrix when dropping the summary candidate from the document cluster. The hypothesis is validated by three theoretical perspectives: semantic scaling, propagation dynamics and matrix perturbation. According to the hypothesis, we formulate the MDS task as the combinatorial optimization of spectral impact and propose an accelerated greedy solution based on a surrogate of spectral impact. The evaluation results on various datasets demonstrate: (1) The performance of the summary candidate is positively correlated with its spectral impact, which accords with our hypothesis; (2) Our spectral-based method has a competitive result as compared to state-of-the-art MDS systems.
While discriminative neural network classifiers are generally preferred, recent work has shown advantages of generative classifiers in term of data efficiency and robustness. In this paper, we focus on natural language inference (NLI). We propose GenNLI, a generative classifier for NLI tasks, and empirically characterize its performance by comparing it to five baselines, including discriminative models and large-scale pretrained language representation models like BERT. We explore training objectives for discriminative fine-tuning of our generative classifiers, showing improvements over log loss fine-tuning from prior work (Lewis and Fan, 2019). In particular, we find strong results with a simple unbounded modification to log loss, which we call the “infinilog loss”. Our experiments show that GenNLI outperforms both discriminative and pretrained baselines across several challenging NLI experimental settings, including small training sets, imbalanced label distributions, and label noise.
In this paper, we focus on the task of fine-grained text sentiment transfer (FGST). This task aims to revise an input sequence to satisfy a given sentiment intensity, while preserving the original semantic content. Different from the conventional sentiment transfer task that only reverses the sentiment polarity (positive/negative) of text, the FTST task requires more nuanced and fine-grained control of sentiment. To remedy this, we propose a novel Seq2SentiSeq model. Specifically, the numeric sentiment intensity value is incorporated into the decoder via a Gaussian kernel layer to finely control the sentiment intensity of the output. Moreover, to tackle the problem of lacking parallel data, we propose a cycle reinforcement learning algorithm to guide the model training. In this framework, the elaborately designed rewards can balance both sentiment transformation and content preservation, while not requiring any ground truth output. Experimental results show that our approach can outperform existing methods by a large margin in both automatic evaluation and human evaluation.
The comprehensive descriptions for factual attribute-value tables, which should be accurate, informative and loyal, can be very helpful for end users to understand the structured data in this form. However previous neural generators might suffer from key attributes missing, less informative and groundless information problems, which impede the generation of high-quality comprehensive descriptions for tables. To relieve these problems, we first propose force attention (FA) method to encourage the generator to pay more attention to the uncovered attributes to avoid potential key attributes missing. Furthermore, we propose reinforcement learning for information richness to generate more informative as well as more loyal descriptions for tables. In our experiments, we utilize the widely used WIKIBIO dataset as a benchmark. Besides, we create WB-filter based on WIKIBIO to test our model in the simulated user-oriented scenarios, in which the generated descriptions should accord with particular user interests. Experimental results show that our model outperforms the state-of-the-art baselines on both automatic and human evaluation.
Automatic story ending generation is an interesting and challenging task in natural language generation. Previous studies are mainly limited to generate coherent, reasonable and diversified story endings, and few works focus on controlling the sentiment of story endings. This paper focuses on generating a story ending which meets the given fine-grained sentiment intensity. There are two major challenges to this task. First is the lack of story corpus which has fine-grained sentiment labels. Second is the difficulty of explicitly controlling sentiment intensity when generating endings. Therefore, we propose a generic and novel framework which consists of a sentiment analyzer and a sentimental generator, respectively addressing the two challenges. The sentiment analyzer adopts a series of methods to acquire sentiment intensities of the story dataset. The sentimental generator introduces the sentiment intensity into decoder via a Gaussian Kernel Layer to control the sentiment of the output. To the best of our knowledge, this is the first endeavor to control the fine-grained sentiment for story ending generation without manually annotating sentiment labels. Experiments show that our proposed framework can generate story endings which are not only more coherent and fluent but also able to meet the given sentiment intensity better.
In this paper, we focus on the task of generating a pun sentence given a pair of word senses. A major challenge for pun generation is the lack of large-scale pun corpus to guide supervised learning. To remedy this, we propose an adversarial generative network for pun generation (Pun-GAN). It consists of a generator to produce pun sentences, and a discriminator to distinguish between the generated pun sentences and the real sentences with specific word senses. The output of the discriminator is then used as a reward to train the generator via reinforcement learning, encouraging it to produce pun sentences which can support two word senses simultaneously. Experiments show that the proposed Pun-GAN can generate sentences that are more ambiguous and diverse in both automatic and human evaluation.
Word Sense Disambiguation (WSD) aims to identify the correct meaning of polysemous words in the particular context. Lexical resources like WordNet which are proved to be of great help for WSD in the knowledge-based methods. However, previous neural networks for WSD always rely on massive labeled data (context), ignoring lexical resources like glosses (sense definitions). In this paper, we integrate the context and glosses of the target word into a unified framework in order to make full use of both labeled data and lexical knowledge. Therefore, we propose GAS: a gloss-augmented WSD neural network which jointly encodes the context and glosses of the target word. GAS models the semantic relationship between the context and the gloss in an improved memory network framework, which breaks the barriers of the previous supervised methods and knowledge-based methods. We further extend the original gloss of word sense via its semantic relations in WordNet to enrich the gloss information. The experimental results show that our model outperforms the state-of-the-art systems on several English all-words WSD datasets.
The goal of Word Sense Disambiguation (WSD) is to identify the correct meaning of a word in the particular context. Traditional supervised methods only use labeled data (context), while missing rich lexical knowledge such as the gloss which defines the meaning of a word sense. Recent studies have shown that incorporating glosses into neural networks for WSD has made significant improvement. However, the previous models usually build the context representation and gloss representation separately. In this paper, we find that the learning for the context and gloss representation can benefit from each other. Gloss can help to highlight the important words in the context, thus building a better context representation. Context can also help to locate the key words in the gloss of the correct word sense. Therefore, we introduce a co-attention mechanism to generate co-dependent representations for the context and gloss. Furthermore, in order to capture both word-level and sentence-level information, we extend the attention mechanism in a hierarchical fashion. Experimental results show that our model achieves the state-of-the-art results on several standard English all-words WSD test datasets.
This paper proposes to study fine-grained coordinated cross-lingual text stream alignment through a novel information network decipherment paradigm. We use Burst Information Networks as media to represent text streams and present a simple yet effective network decipherment algorithm with diverse clues to decipher the networks for accurate text stream alignment. Experiments on Chinese-English news streams show our approach not only outperforms previous approaches on bilingual lexicon extraction from coordinated text streams but also can harvest high-quality alignments from large amounts of streaming data for endless language knowledge mining, which makes it promising to be a new paradigm for automatic language knowledge acquisition.
Multi-document summarization provides users with a short text that summarizes the information in a set of related documents. This paper introduces affinity-preserving random walk to the summarization task, which preserves the affinity relations of sentences by an absorbing random walk model. Meanwhile, we put forward adjustable affinity-preserving random walk to enforce the diversity constraint of summarization in the random walk process. The ROUGE evaluations on DUC 2003 topic-focused summarization task and DUC 2004 generic summarization task show the good performance of our method, which has the best ROUGE-2 recall among the graph-based ranking methods.
Distant-supervised relation extraction inevitably suffers from wrong labeling problems because it heuristically labels relational facts with knowledge bases. Previous sentence level denoise models don’t achieve satisfying performances because they use hard labels which are determined by distant supervision and immutable during training. To this end, we introduce an entity-pair level denoise method which exploits semantic information from correctly labeled entity pairs to correct wrong labels dynamically during training. We propose a joint score function which combines the relational scores based on the entity-pair representation and the confidence of the hard label to obtain a new label, namely a soft label, for certain entity pair. During training, soft labels instead of hard labels serve as gold labels. Experiments on the benchmark dataset show that our method dramatically reduces noisy instances and outperforms other state-of-the-art systems.
Previous studies on Chinese semantic role labeling (SRL) have concentrated on a single semantically annotated corpus. But the training data of single corpus is often limited. Whereas the other existing semantically annotated corpora for Chinese SRL are scattered across different annotation frameworks. But still, Data sparsity remains a bottleneck. This situation calls for larger training datasets, or effective approaches which can take advantage of highly heterogeneous data. In this paper, we focus mainly on the latter, that is, to improve Chinese SRL by using heterogeneous corpora together. We propose a novel progressive learning model which augments the Progressive Neural Network with Gated Recurrent Adapters. The model can accommodate heterogeneous inputs and effectively transfer knowledge between them. We also release a new corpus, Chinese SemBank, for Chinese SRL. Experiments on CPB 1.0 show that our model outperforms state-of-the-art methods.
Knowledge graph (KG) completion adds new facts to a KG by making inferences from existing facts. Most existing methods ignore the time information and only learn from time-unknown fact triples. In dynamic environments that evolve over time, it is important and challenging for knowledge graph completion models to take into account the temporal aspects of facts. In this paper, we present a novel time-aware knowledge graph completion model that is able to predict links in a KG using both the existing facts and the temporal information of the facts. To incorporate the happening time of facts, we propose a time-aware KG embedding model using temporal order information among facts. To incorporate the valid time of facts, we propose a joint time-aware inference model based on Integer Linear Programming (ILP) using temporal consistencyinformationasconstraints. Wefurtherintegratetwomodelstomakefulluseofglobal temporal information. We empirically evaluate our models on time-aware KG completion task. Experimental results show that our time-aware models achieve the state-of-the-art on temporal facts consistently.
Recognizing Textual Entailment (RTE) is a fundamentally important task in natural language processing that has many applications. The recently released Stanford Natural Language Inference (SNLI) corpus has made it possible to develop and evaluate deep neural network methods for the RTE task. Previous neural network based methods usually try to encode the two sentences (premise and hypothesis) and send them together into a multi-layer perceptron to get their entailment type, or use LSTM-RNN to link two sentences together while using attention mechanic to enhance the model’s ability. In this paper, we propose to use the re-read mechanic, which means to read the premise again and again while reading the hypothesis. After read the premise again, the model can get a better understanding of the premise, which can also affect the understanding of the hypothesis. On the contrary, a better understanding of the hypothesis can also affect the understanding of the premise. With the alternative re-read process, the model can “think” of a better decision of entailment type. We designed a new LSTM unit called re-read LSTM (rLSTM) to implement this “thinking” process. Experiments show that we achieve results better than current state-of-the-art equivalents.
Retrospective event detection is an important task for discovering previously unidentified events in a text stream. In this paper, we propose two fast centroid-aware event detection models based on a novel text stream representation – Burst Information Networks (BINets) for addressing the challenge. The BINets are time-aware, efficient and can be easily analyzed for identifying key information (centroids). These advantages allow the BINet-based approaches to achieve the state-of-the-art performance on multiple datasets, demonstrating the efficacy of BINets for the task of event detection.
Algorithms for automatic term extraction in a specific domain should consider at least two issues, namely Unithood and Termhood (Kageura, 1996). Unithood refers to the degree of a string to occur as a word or a phrase. Termhood (Chen Yirong, 2005) refers to the degree of a word or a phrase to occur as a domain specific concept. Unlike unithood, study on termhood is not yet widely reported. In classified corpora, the class information provides the cue to the nature of data and can be used in termhood calculation. Three algorithms are provided and evaluated to investigate termhood based on classified corpora. The three algorithms are based on lexicon set computing, term frequency and document frequency, and the strength of the relation between a term and its document class respectively. Our objective is to investigate the effects of these different termhood measurement features. After evaluation, we can find which features are more effective and also, how we can improve these different features to achieve the best performance. Preliminary results show that the first measure can effectively filter out independent terms or terms of general use.