Croatia

Modern fact verification systems have distanced themselves from the black box paradigm by providing the evidence used to infer their veracity judgments. Hence, evidence-backed fact verification systems&#39; performance heavily depends on the capabilities of their retrieval component to identify these facts. A popular evaluation benchmark for these systems is the FEVER task, which consists of determining the veracity of short claims using sentences extracted from Wikipedia. In this paper, we present a novel approach to the the retrieval steps of the FEVER task leveraging the graph structure of Wikipedia. The retrieval models surpass state of the art results at both sentence and document level. Additionally, we show that by feeding our retrieved evidence to the best-performing textual entailment model, we set a new state of the art in the FEVER competition.

EACL 2023

Enhancing Information Retrieval in Fact Extraction and Verification

fever task

document retrieval

fact verification

information retrieval

Modern fact verification systems have distanced themselves from the black box paradigm by providing the evidence used to infer their veracity judgments. Hence, evidence-backed fact verification systems' performance heavily depends on the capabilities of their retrieval component to identify these facts. A popular evaluation benchmark for these systems is the FEVER task, which consists of determining the veracity of short claims using sentences extracted from Wikipedia. In this paper, we present a novel approach to the the retrieval steps of the FEVER task leveraging the graph structure of Wikipedia. The retrieval models surpass state of the art results at both sentence and document level. Additionally, we show that by feeding our retrieved evidence to the best-performing textual entailment model, we set a new state of the art in the FEVER competition.

workshop paper

### Welcome to the 17th Conference of the European Chapter of the Association for Computational Linguistics (EACL) 
Continuing its mission of expanding and involving the science community of all European countries, EACL had selected the Ukrainian community for the 16th EACL, which took place online due to the COVID pandemic. Unfortunately, the ongoing war made the organisation in Kyiv impossible. Considering the importance of physical interaction among researchers, especially after the restrictions imposed by the COVID pandemic, in addition to an online mode, the conference will be held in Dubrovnik, Croatia, from 2 to 6 of May, 2023. The original aim of strengthening the connection with the Ukrainian community will not change as our program will feature a dedicated session and a workshop to highlight work on Ukrainian language technologies. As the flagship European conference in the field of computational linguistics, EACL welcomes European and international researchers covering a broad spectrum of research areas that are concerned with computational approaches to natural language.

You need to register for the conference in order to access this site. Please visit https://2023.eacl.org/registration for more information.

As the flagship European conference in the field of computational linguistics, EACL welcomes European and international researchers covering a broad spectrum of research areas that are concerned with computational approaches to natural language.

Recently it has been shown that without any access to the contextual passage, multiple choice reading comprehension (MCRC) systems are able to answer questions significantly better than random on average. These systems use their accumulated "world knowledge" to directly answer questions, rather than using information from the passage. This paper examines the possibility of exploiting this observation as a tool for test designers to ensure that the form of "world knowledge" is acceptable for a particular set of questions. We propose information-theory based metrics that enable the level of "world knowledge" exploited by systems to be assessed. Two metrics are described: the expected number of options, which measures whether a passage-free system can identify the answer a question using world knowledge; and the contextual mutual information, which measures the importance of context for a given question. We demonstrate that questions with low expected number of options, and hence answerable by the shortcut system, are often similarly answerable by humans without context. This highlights that the general knowledge 'shortcuts' could be equally used by exam candidates, and that our proposed metrics may be helpful for future test designers to monitor the quality of questions.

"World Knowledge" in Multiple Choice Reading Comprehension

Automatic fact verification has become an increasingly popular topic in recent years and among datasets the Fact Extraction and VERification (FEVER) dataset is one of the most popular. In this work we present BEVERS, a tuned baseline system for the FEVER dataset. Our pipeline uses standard approaches for document retrieval, sentence selection, and final claim classification, however, we spend considerable effort ensuring optimal performance for each component. The results are that BEVERS achieves the highest FEVER score and label accuracy among all systems, published or unpublished. We also apply this pipeline to another fact verification dataset, Scifact, and achieve the highest label accuracy among all systems on that dataset as well. We also make our full code available.

BEVERS: A General, Simple, and Performant Framework for Automatic Fact Verification

Explanation prompts ask language models to not only assign a label to a given input, such as entailment or contradiction in natural language inference (NLI) tasks, but also to generate a free-text explanation that supports this label. While explanation prompts were originally introduced aiming to improve model interpretability, here we show that they also improve robustness to superficial cues. Compared to prompting for labels only, explanation prompt- ing shows stronger performance on adversarial NLI benchmarks, outperforming state of the art on ANLI, Counterfactually-Augmented NLI, and SNLI-Hard datasets. Analysis suggests that the increase in robustness is due to a reduction in the association strength be- tween single tokens and labels, i.e., explanation prompting weakens superficial cues. More specifically, we find that single tokens that are highly predictive of the correct answer in the label-only setting become uninformative when the model also has to generate explanations.

Prompting for explanations improves Adversarial NLI. Is this true? {Yes} it is {true} because {it weakens superficial cues}

In my talk I will discuss bias recognition on BASIL, a dataset of political news which has been annotated with different kinds of biases. The start of the presentation unveils several inconsistencies in prior work using this dataset, showing that most approaches focus only on certain task formulations while ignoring others, and also failing to report important evaluation details. Later I discuss a comprehensive categorization of these approaches, as well as a more uniform and clear set of evaluation metrics. In the end, the importance of the missing formulations and propose the novel task of simultaneously detecting different kinds of biases in news is discussed and various task formulations on six different BASIL classification tasks provide a unified manner for detecting bias in general.

An Effective Approach for Informational and Lexical Bias Detection

Leveraging contextual knowledge has become standard practice in automated claim verification, yet the impact of temporal reasoning has been largely overlooked. Our study demonstrates that time positively influences the claim verification process of evidence-based factchecking. The temporal aspects and relations between claims and evidence are first established through grounding on shared timelines, which are constructed using publication dates and time expressions extracted from their
text. Temporal information is then provided to RNN-based and Transformer-based classifiers before or after claim and evidence encoding. Our time-aware fact-checking models surpass base models by up to 9% Micro F1 (64.17%) and 15% Macro F1 (47.43%) on the MultiFC dataset. They also outperform prior methods that explicitly model temporal relations between evidence. Our findings show that the presence of temporal information and the manner in which timelines are constructed greatly influence how fact-checking models determine the relevance and supporting or refuting character of evidence documents.

Implicit Temporal Reasoning for Evidence-Based Fact-Checking

Many computational argumentation tasks, such as stance classification, are topic-dependent: The effectiveness of approaches to these tasks depends on whether they are trained on the same topics as those on which they are tested. The key question is: What are these training topics? To answer this question, we take the first step of mapping the argumentation landscape with an argument topic ontology which is build on the World Economic Forum, Wikipedia's list of controversial topics, and Debatepedia. By comparing the topics in our ontology with those in 59 argument corpora, we perform the first comprehensive assessment of their topic coverage. Our study shows that existing argument corpora barely cover topics in our topic ontology. This points to a new goal for corpus construction to achieve a broad topic coverage and thus better generalizability of computational argumentation approaches.

Topic Ontologies for Arguments

Fake news and misinformation spread rapidly on the Internet. How to identify it and how to interpret the identification results have become important issues. In this paper, we propose a Dual Co-Attention Network (Dual-CAN) for fake news detection, which takes news con- tent, social media replies, and external knowl- edge into consideration. Our experimental re- sults support that the proposed Dual-CAN out- performs current representative models in two benchmark datasets. We further make in-depth discussions by comparing how models work in both datasets with empirical analysis of atten- tion weights.

Entity-Aware Dual Co-Attention Network for Fake News Detection

Contrastive pre-training on distant supervision has shown remarkable effectiveness in improving supervised relation extraction tasks. However, the existing methods ignore the intrinsic noise of distant supervision during the pre-training stage. In this paper, we propose a weighted contrastive learning method by leveraging the supervised data to estimate the reliability of pre-training instances and explicitly reduce the effect of noise. Experimental results on three supervised datasets demonstrate the advantages of our proposed weighted contrastive learning approach compared to two state-of-the-art non-weighted baselines.

Relation Extraction with Weighted Contrastive Pre-training on Distant Supervision

Language models are trained on large volumes of text, and as a result their parameters might contain a significant body of factual knowledge. Any downstream task performed by these models implicitly builds on these facts, and thus it is highly desirable to have means for representing this body of knowledge in an interpretable way. However, there is currently no mechanism for such a representation. Here, we propose to address this goal by extracting a knowledge-graph of facts from a given language model. We describe a procedure for "crawling'' the internal knowledge-base of a language model. Specifically, given a seed entity, we expand a knowledge-graph around it. The crawling procedure is decomposed into sub-tasks, realized through specially designed prompts that control for both precision (i.e., that no wrong facts are generated) and recall (i.e., the number of facts generated). We evaluate our approach on graphs crawled starting from dozens of seed entities, and show it yields high precision graphs (82-92%), while emitting a reasonable number of facts per entity.

Crawling The Internal Knowledge-Base of Language Models

Recent studies have suggested that neural language models learn and store a large amount of facts and commonsense knowledge from training data. The ability of language models to restore such knowledge is often evaluated via zero-shot cloze-style QA tasks. However, such evaluations rely only on prediction accuracy without punishing the systems for their mistakes, e.g., simply guessing or hallucinating likely answers. Selective prediction is a more informative evaluation framework that takes the confidence of predictions into account. Under the selective prediction setting, a model is evaluated not only by the number of correct predictions, but also by the ability to filter out dubious predictions by estimating the confidence of individual predictions. Such confidence-aware evaluation is crucial for determining whether to
trust zero-shot predictions of language models. In this paper, we apply the selective prediction setting to an existing benchmark, LAMA probe, and conduct extensive experiments with recent neural language models and different confidence functions. We empirically show that our Selective-LAMA evaluation is more robust to the effect of simple guesses than the conventional accuracy-based evaluation. Our
evaluation reveals the importance of the choice of confidence functions by showing that simply relying on token probabilities is not always the best choice. Further analysis shows that various confidence functions exhibit different preferences over predicted tokens for a given context.

Downloads

Next from EACL 2023

"World Knowledge" in Multiple Choice Reading Comprehension

.css-70qvj9{display:-webkit-box;display:-webkit-flex;display:-ms-flexbox;display:flex;-webkit-align-items:center;-webkit-box-align:center;-ms-flex-align:center;align-items:center;}Downloads

Next from EACL 2023

"World Knowledge" in Multiple Choice Reading Comprehension

Downloads