Automated fact-checking systems verify claims against evidence to predict their veracity. In real-world scenarios, the retrieved evidence may not unambiguously support or refute the claim and yield conflicting but valid interpretations. Existing fact-checking datasets assume that the models developed with them predict a single veracity label for each claim, thus discouraging the handling of such ambiguity. To address this issue we present AmbiFC,1 a fact-checking dataset with 10k claims derived from real-world information needs. It contains fine-grained evidence annotations of 50k passages from 5k Wikipedia pages. We analyze the disagreements arising from ambiguity when comparing claims against evidence in AmbiFC, observing a strong correlation of annotator disagreement with linguistic phenomena such as underspecification and probabilistic reasoning. We develop models for predicting veracity handling this ambiguity via soft labels, and find that a pipeline that learns the label distribution for sentence-level evidence selection and veracity prediction yields the best performance. We compare models trained on different subsets of AmbiFC and show that models trained on the ambiguous instances perform better when faced with the identified linguistic phenomena.
Generative retrieval models encode pointers to information in a corpus as an index within the model’s parameters. These models serve as part of a larger pipeline, where retrieved information conditions generation for knowledge-intensive NLP tasks. However, we identify two limitations: the generative retrieval does not account for contextual information. Secondly, the retrieval can’t be tuned for the downstream readers as decoding the page title is a non-differentiable operation. This paper introduces Re3val, trained with generative reranking and reinforcement learning using limited data. Re3val leverages context acquired via Dense Passage Retrieval to rerank the retrieved page titles and utilizes REINFORCE to maximize rewards generated by constrained decoding. Additionally, we generate questions from our pre-training dataset to mitigate epistemic uncertainty and bridge the domain gap between the pre-training and fine-tuning datasets. Subsequently, we extract and rerank contexts from the KILT database using the rerank page titles. Upon grounding the top five reranked contexts, Re3val demonstrates the Top 1 KILT scores compared to all other generative retrieval models across five KILT datasets.
Deriving meaningful sentence embeddings is crucial in capturing the semantic relationship between texts. Recent advances in building sentence embedding models have centered on replacing traditional human-generated text datasets with those generated by LLMs. However, the properties of these widely used LLM-generated texts remain largely unexplored. Here, we evaluate the quality of the LLM-generated texts from four perspectives (Positive Text Repetition, Length Difference Penalty, Positive Score Compactness, and Negative Text Implausibility) and find that there exists an inherent difference between human and LLM-generated datasets. To further enhance sentence embeddings using both human and LLM-generated datasets, we propose a novel loss function that incorporates Positive-Negative sample Augmentation (PNA) within the contrastive learning objective. Our results demonstrate that PNA effectively mitigates the sentence anisotropy problem in Wikipedia corpus (-7% compared to CLHAIF) and simultaneously improves the Spearman’s correlation in standard Semantic Textual Similarity (STS) tasks (+1.47% compared to CLHAIF).
In this study, we introduce BEnQA, a dataset comprising parallel Bengali and English exam questions for middle and high school levels in Bangladesh. Our dataset consists of approximately 5K questions covering several subjects in science with different types of questions, including factual, application, and reasoning-based questions. We benchmark several Large Language Models (LLMs) with our parallel dataset and observe a notable performance disparity between the models in Bengali and English. We also investigate some prompting methods, and find that Chain-of-Thought prompting is beneficial mostly on reasoning questions, but not so much on factual ones. We also find that appending English translation helps to answer questions in Bengali. Our findings point to promising future research directions for improving the performance of LLMs in Bengali and more generally in low-resource languages.
This paper investigates the inherent knowledge in language models from the perspective of epistemological holism. The purpose of this paper is to explore whether LLMs exhibit characteristics consistent with epistemological holism. These characteristics suggest that core knowledge, such as commonsense, general, and specific knowledge, each plays a specific role, serving as the foundation of our knowledge system and being difficult to revise. To assess these traits related to holism, we created a scientific reasoning dataset and examined the epistemology of language models through three tasks: Abduction, Revision, and Argument Generation. In the abduction task, the language models explained situations while avoiding revising the core knowledge. However, in other tasks, the language models were revealed not to distinguish between core and peripheral knowledge, showing an incomplete alignment with holistic knowledge principles.
Application of LLM to database queries on natural language sentences has demonstrated impressive results in both single and multi-hop scenarios.In the existing methodologies, the requirement to re-encode query vectors at each stage for processing multi-hop queries presents a significant bottleneck to the inference speed.This paper proposes VKGFR (Virtual Knowledge Graph based Fact Retriever) that leverages large language models to extract representations corresponding to a sentence’s knowledge graph, significantly enhancing inference speed for multi-hop reasoning without performance loss.Given that both the queries and natural language database sentences can be structured as a knowledge graph, we suggest extracting a Virtual Knowledge Graph (VKG) representation from sentences with LLM.Over the pre-constructed VKG, our VKGFR conducts retrieval with a tiny model structure, showing performance improvements with higher computational efficiency. We evaluate VKGFR on the WikiNLDB and MetaQA dataset, designed for multi-hop database reasoning over text. The results indicate 13x faster inference speed on the WikiNLDB dataset without performance loss.
This paper reports on the argument mining system submitted to the ArgMining workshop 2024 for The Perspective Argument Retrieval Shared Task (Falk et al., 2024). We com- bine the strengths of a smaller Sentence BERT model and a Large Language Model: the for- mer is fine-tuned for a contrastive embedding objective and a classification objective whereas the latter is invoked to augment the query and populate the latent space with diverse relevant arguments. We conduct an ablation study on these components to find that each contributes substantially to the diversity and relevance cri- teria for the top-k retrieval of arguments from the given corpus.
Despite the rapid development of large language models (LLMs) for the Korean language, there remains an obvious lack of benchmark datasets that test the requisite Korean cultural and linguistic knowledge. Because many existing Korean benchmark datasets are derived from the English counterparts through translation, they often overlook the different cultural contexts. For the few benchmark datasets that are sourced from Korean data capturing cultural knowledge, only narrow tasks such as hate speech detection are offered. To address this gap, we introduce a benchmark of Cultural and Linguistic Intelligence in Korean (CLIcK), a dataset comprising 1,995 QA pairs. CLIcK sources its data from official Korean exams and textbooks, partitioning the questions into eleven categories under the two main categories of language and culture. For each instance in click, we provide fine-grained annotation of which cultural and linguistic knowledge is required to correctly answer the question. Using CLIcK, we test 13 language models to assess their performance. Our evaluation uncovers insights into their performances across the categories, as well as the diverse factors affecting their comprehension. CLIcK offers the first large-scale comprehensive Korean-centric analysis of LLMs’ proficiency in Korean language and culture.
In real world applications, knowledge graphs (KG) are widely used in various domains (e.g. medical applications and dialogue agents). However, for fact verification, KGs have not been adequately utilized as a knowledge source. KGs can be a valuable knowledge source in fact verification due to their reliability and broad applicability. A KG consists of nodes and edges which makes it clear how concepts are linked together, allowing machines to reason over chains of topics. However, there are many challenges in understanding how these machine-readable concepts map to information in text. To enable the community to better use KGs, we introduce a new dataset, FactKG: Fact Verificationvia Reasoning on Knowledge Graphs. It consists of 108k natural language claims with five types of reasoning: One-hop, Conjunction, Existence, Multi-hop, and Negation. Furthermore, FactKG contains various linguistic patterns, including colloquial style claims as well as written style claims to increase practicality. Lastly, we develop a baseline approach and analyze FactKG over these reasoning types. We believe FactKG can advance both reliability and practicality in KG-based fact verification.
With the proliferation of social media, accurate detection of hate speech has become critical to ensure safety online. To combat nuanced forms of hate speech, it is important to identify and thoroughly explain hate speech to help users understand its harmful effects. Recent benchmarks have attempted to tackle this issue by training generative models on free-text annotations of implications in hateful text. However, we find significant reasoning gaps in the existing annotations schemes, which may hinder the supervision of detection models. In this paper, we introduce a hate speech detection framework, **HARE**, which harnesses the reasoning capabilities of large language models (LLMs) to fill these gaps in explanations of hate speech, thus enabling effective supervision of detection models. Experiments on SBIC and Implicit Hate benchmarks show that our method, using model-generated data, consistently outperforms baselines, using existing free-text human annotations. Analysis demonstrates that our method enhances the explanation quality of trained models and improves generalization to unseen datasets. Our code is available at https://github.com/joonkeekim/hare-hate-speech.git.
We address an important gap in detecting political bias in news articles. Previous works that perform document classification can be influenced by the writing style of each news outlet, leading to overfitting and limited generalizability. Our approach overcomes this limitation by considering both the sentence-level semantics and the document-level rhetorical structure, resulting in a more robust and style-agnostic approach to detecting political bias in news articles. We introduce a novel multi-head hierarchical attention model that effectively encodes the structure of long documents through a diverse ensemble of attention heads. While journalism follows a formalized rhetorical structure, the writing style may vary by news outlet. We demonstrate that our method overcomes this domain dependency and outperforms previous approaches for robustness and accuracy. Further analysis and human evaluation demonstrate the ability of our model to capture common discourse structures in journalism.
Recent works in open-domain question answering (QA) have explored generating context passages from large language models (LLMs), replacing the traditional retrieval step in the QA pipeline. However, it is not well understood why generated passages can be more effective than retrieved ones. This study revisits the conventional formulation of QA and introduces the concept of knowledge corpus error. This error arises when the knowledge corpus used for retrieval is only a subset of the entire string space, potentially excluding more helpful passages that exist outside the corpus. LLMs may mitigate this shortcoming by generating passages in a larger space. We come up with an experiment of paraphrasing human-annotated gold context using LLMs to observe knowledge corpus error empirically. Our results across three QA benchmarks reveal an increased performance (10% - 13%) when using paraphrased passage, indicating a signal for the existence of knowledge corpus error.
For knowledge intensive NLP tasks, it has been widely accepted that accessing more information is a contributing factor to improvements in the model’s end-to-end performance. However, counter-intuitively, too much context can have a negative impact on the model when evaluated on common question answering (QA) datasets. In this paper, we analyze how passages can have a detrimental effect on retrieve-then-read architectures used in question answering. Our empirical evidence indicates that the current read architecture does not fully leverage the retrieved passages and significantly degrades its performance when using the whole passages compared to utilizing subsets of them. Our findings demonstrate that model accuracy can be improved by 10% on two popular QA datasets by filtering out detrimental passages. Additionally, these outcomes are attained by utilizing existing retrieval methods without further training or data. We further highlight the challenges associated with identifying the detrimental passages. First, even with the correct context, the model can make an incorrect prediction, posing a challenge in determining which passages are most influential. Second, evaluation typically considers lexical matching, which is not robust to variations of correct answers. Despite these limitations, our experimental results underscore the pivotal role of identifying and removing these detrimental passages for the context-efficient retrieve-then-read pipeline.
Large language models (LLMs) have shown impressive achievements in solving a broad range of tasks. Augmented by instruction fine-tuning, LLMs have also been shown to generalize in zero-shot settings as well. However, whether LLMs closely align with the human disagreement distribution has not been well-studied, especially within the scope of natural language inference (NLI). In this paper, we evaluate the performance and alignment of LLM distribution with humans using two different techniques to estimate the multinomial distribution: Monte Carlo Estimation (MCE) and Log Probability Estimation (LPE). As a result, we show LLMs exhibit limited ability in solving NLI tasks and simultaneously fail to capture human disagreement distribution. The inference and human alignment performances plunge even further on data samples with high human disagreement levels, raising concerns about their natural language understanding (NLU) ability and their representativeness to a larger human population.
Document retrieval is a core component of many knowledge-intensive natural language processing task formulations such as fact verification. Sources of textual knowledge such as Wikipedia articles condition the generation of answers from the models. Recent advances in retrieval use sequence-to-sequence models to incrementally predict the title of the appropriate Wikipedia page given an input instance. However, this method requires supervision in the form of human annotation to label which Wikipedia pages contain appropriate context. This paper introduces a distant-supervision method that does not require any annotation train auto-regressive retrievers that attain competitive R-Precision and Recall in a zero-shot setting. Furthermore we show that with task-specific supervised fine-tuning, auto-regressive retrieval performance for two Wikipedia-based fact verification tasks can approach or even exceed full supervision using less than 1/4 of the annotated data. We release all code and models
Neural models have shown impressive performance gains in answering queries from natural language text. However, existing works are unable to support database queries, such as “List/Count all female athletes who were born in 20th century”, which require reasoning over sets of relevant facts with operations such as join, filtering and aggregation. We show that while state-of-the-art transformer models perform very well for small databases, they exhibit limitations in processing noisy data, numerical operations, and queries that aggregate facts. We propose a modular architecture to answer these database-style queries over multiple spans from text and aggregating these at scale. We evaluate the architecture using WikiNLDB, a novel dataset for exploring such queries. Our architecture scales to databases containing thousands of facts whereas contemporary models are limited by how many facts can be encoded. In direct comparison on small databases, our approach increases overall answer accuracy from 85% to 90%. On larger databases, our approach retains its accuracy whereas transformer baselines could not encode the context.
This paper introduces the task of factual error correction: performing edits to a claim so that the generated rewrite is better supported by evidence. This extends the well-studied task of fact verification by providing a mechanism to correct written texts that are refuted or only partially supported by evidence. We demonstrate that it is feasible to train factual error correction systems from existing fact checking datasets which only contain labeled claims accompanied by evidence, but not the correction. We achieve this by employing a two-stage distant supervision approach that incorporates evidence into masked claims when generating corrections. Our approach, based on the T5 transformer and using retrieved evidence, achieved better results than existing work which used a pointer copy network and gold evidence, producing accurate factual error corrections for 5x more instances in human evaluation and a .125 increase in SARI score. The evaluation is conducted on a dataset of 65,000 instances based on a recent fact verification shared task and we release it to enable further work on the task.
Challenging problems such as open-domain question answering, fact checking, slot filling and entity linking require access to large, external knowledge sources. While some models do well on individual tasks, developing general models is difficult as each task might require computationally expensive indexing of custom knowledge sources, in addition to dedicated infrastructure. To catalyze research on models that condition on specific information in large textual resources, we present a benchmark for knowledge-intensive language tasks (KILT). All tasks in KILT are grounded in the same snapshot of Wikipedia, reducing engineering turnaround through the re-use of components, as well as accelerating research into task-agnostic memory architectures. We test both task-specific and general baselines, evaluating downstream performance in addition to the ability of the models to provide provenance. We find that a shared dense vector index coupled with a seq2seq model is a strong baseline, outperforming more tailor-made approaches for fact checking, open-domain question answering and dialogue, and yielding competitive results on entity linking and slot filling, by generating disambiguated text. KILT data and code are available at https://github.com/facebookresearch/KILT.
The Fact Extraction and VERification Over Unstructured and Structured information (FEVEROUS) shared task, asks participating systems to determine whether human-authored claims are Supported or Refuted based on evidence retrieved from Wikipedia (or NotEnoughInfo if the claim cannot be verified). Compared to the FEVER 2018 shared task, the main challenge is the addition of structured data (tables and lists) as a source of evidence. The claims in the FEVEROUS dataset can be verified using only structured evidence, only unstructured evidence, or a mixture of both. Submissions are evaluated using the FEVEROUS score that combines label accuracy and evidence retrieval. Unlike FEVER 2018, FEVEROUS requires partial evidence to be returned for NotEnoughInfo claims, and the claims are longer and thus more complex. The shared task received 13 entries, six of which were able to beat the baseline system. The winning team was “Bust a move!”, achieving a FEVEROUS score of 27% (+9% compared to the baseline). In this paper we describe the shared task, present the full results and highlight commonalities and innovations among the participating systems.
The biases present in training datasets have been shown to affect models for sentence pair classification tasks such as natural language inference (NLI) and fact verification. While fine-tuning models on additional data has been used to mitigate them, a common issue is that of catastrophic forgetting of the original training dataset. In this paper, we show that elastic weight consolidation (EWC) allows fine-tuning of models to mitigate biases while being less susceptible to catastrophic forgetting. In our evaluation on fact verification and NLI stress tests, we show that fine-tuning with EWC dominates standard fine-tuning, yielding models with lower levels of forgetting on the original (biased) dataset for equivalent gains in accuracy on the fine-tuning (unbiased) dataset.
The task of Natural Language Inference (NLI) is widely modeled as supervised sentence pair classification. While there has been a lot of work recently on generating explanations of the predictions of classifiers on a single piece of text, there have been no attempts to generate explanations of classifiers operating on pairs of sentences. In this paper, we show that it is possible to generate token-level explanations for NLI without the need for training data explicitly annotated for this purpose. We use a simple LSTM architecture and evaluate both LIME and Anchor explanations for this task. We compare these to a Multiple Instance Learning (MIL) method that uses thresholded attention make token-level predictions. The approach we present in this paper is a novel extension of zero-shot single-sentence tagging to sentence pairs for NLI. We conduct our experiments on the well-studied SNLI dataset that was recently augmented with manually annotation of the tokens that explain the entailment relation. We find that our white-box MIL-based method, while orders of magnitude faster, does not reach the same accuracy as the black-box methods.
Automated fact verification has been progressing owing to advancements in modeling and availability of large datasets. Due to the nature of the task, it is critical to understand the vulnerabilities of these systems against adversarial instances designed to make them predict incorrectly. We introduce two novel scoring metrics, attack potency and system resilience which take into account the correctness of the adversarial instances, an aspect often ignored in adversarial evaluations. We consider six fact verification systems from the recent Fact Extraction and VERification (FEVER) challenge: the four best-scoring ones and two baselines. We evaluate adversarial instances generated by a recently proposed state-of-the-art method, a paraphrasing method, and rule-based attacks devised for fact verification. We find that our rule-based attacks have higher potency, and that while the rankings among the top systems changed, they exhibited higher resilience than the baselines.
We present the results of the second Fact Extraction and VERification (FEVER2.0) Shared Task. The task challenged participants to both build systems to verify factoid claims using evidence retrieved from Wikipedia and to generate adversarial attacks against other participant’s systems. The shared task had three phases: building, breaking and fixing. There were 8 systems in the builder’s round, three of which were new qualifying submissions for this shared task, and 5 adversaries generated instances designed to induce classification errors and one builder submitted a fixed system which had higher FEVER score and resilience than their first submission. All but one newly submitted systems attained FEVER scores higher than the best performing system from the first shared task and under adversarial evaluation, all systems exhibited losses in FEVER score. There was a great variety in adversarial attack types as well as the techniques used to generate the attacks, In this paper, we present the results of the shared task and a summary of the systems, highlighting commonalities and innovations among participating systems.
In this paper we introduce a new publicly available dataset for verification against textual sources, FEVER: Fact Extraction and VERification. It consists of 185,445 claims generated by altering sentences extracted from Wikipedia and subsequently verified without knowledge of the sentence they were derived from. The claims are classified as Supported, Refuted or NotEnoughInfo by annotators achieving 0.6841 in Fleiss kappa. For the first two classes, the annotators also recorded the sentence(s) forming the necessary evidence for their judgment. To characterize the challenge of the dataset presented, we develop a pipeline approach and compare it to suitably designed oracles. The best accuracy we achieve on labeling a claim accompanied by the correct evidence is 31.87%, while if we ignore the evidence we achieve 50.91%. Thus we believe that FEVER is a challenging testbed that will help stimulate progress on claim verification against textual sources.
The recently increased focus on misinformation has stimulated research in fact checking, the task of assessing the truthfulness of a claim. Research in automating this task has been conducted in a variety of disciplines including natural language processing, machine learning, knowledge representation, databases, and journalism. While there has been substantial progress, relevant papers and articles have been published in research communities that are often unaware of each other and use inconsistent terminology, thus impeding understanding and further progress. In this paper we survey automated fact checking research stemming from natural language processing and related disciplines, unifying the task formulations and methodologies across papers and authors. Furthermore, we highlight the use of evidence as an important distinguishing factor among them cutting across task formulations and methods. We conclude with proposing avenues for future NLP research on automated fact checking.
We present the results of the first Fact Extraction and VERification (FEVER) Shared Task. The task challenged participants to classify whether human-written factoid claims could be SUPPORTED or REFUTED using evidence retrieved from Wikipedia. We received entries from 23 competing teams, 19 of which scored higher than the previously published baseline. The best performing system achieved a FEVER score of 64.21%. In this paper, we present the results of the shared task and a summary of the systems, highlighting commonalities and innovations among participating systems.
In this paper we present our automated fact checking system demonstration which we developed in order to participate in the Fast and Furious Fact Check challenge. We focused on simple numerical claims such as “population of Germany in 2015 was 80 million” which comprised a quarter of the test instances in the challenge, achieving 68% accuracy. Our system extends previous work on semantic parsing and claim identification to handle temporal expressions and knowledge bases consisting of multiple tables, while relying solely on automatically generated training data. We demonstrate the extensible nature of our system by evaluating it on relations used in previous work. We make our system publicly available so that it can be used and extended by the community.
Fake news has become a hotly debated topic in journalism. In this paper, we present our entry to the 2017 Fake News Challenge which models the detection of fake news as a stance classification task that finished in 11th place on the leader board. Our entry is an ensemble system of classifiers developed by students in the context of their coursework. We show how we used the stacking ensemble method for this purpose and obtained improvements in classification accuracy exceeding each of the individual models’ performance on the development data. Finally, we discuss aspects of the experimental setup of the challenge.