Hypothetical Documents or Knowledge Leakage? Rethinking LLM-based Query Expansion

Query expansion methods powered by large language models (LLMs) have demonstrated effectiveness in zero-shot retrieval tasks. These methods assume that LLMs can generate hypothetical documents that, when incorporated into a query vector, enhance the retrieval of real evidence. However, we challenge this assumption by investigating whether knowledge leakage in benchmarks contributes to the observed performance gains. Using fact verification as a testbed, we analyzed whether the generated documents contained information entailed by ground truth evidence and assessed their impact on performance. Our findings indicate that performance improvements occurred consistently only for claims whose generated documents included sentences entailed by ground truth evidence. This suggests that knowledge leakage may be present in these benchmarks, inflating the perceived performance of LLM-based query expansion methods, particularly in real-world scenarios that require retrieving niche or novel knowledge.

HerO at AVeriTeC: The Herd of Open Large Language Models for Verifying Real-World Claims

To tackle the AVeriTeC shared task hosted by the FEVER-24, we introduce a system that only employs publicly available large language models (LLMs) for each step of automated fact-checking, dubbed the Herd of Open LLMs for verifying real-world claims (HerO). HerO employs multiple LLMs for each step of automated fact-checking. For evidence retrieval, a language model is used to enhance a query by generating hypothetical fact-checking documents. We prompt pretrained and fine-tuned LLMs for question generation and veracity prediction by crafting prompts with retrieved in-context samples. HerO achieved 2nd place on the leaderboard with the AVeriTeC score of 0.57, suggesting the potential of open LLMs for verifying real-world claims. For future research, we make our code publicly available at https://github.com/ssu-humane/HerO.

ChatGPT Rates Natural Language Explanation Quality Like Humans: But on Which Scales?

As AI becomes more integral in our lives, the need for transparency and responsibility grows. While natural language explanations (NLEs) are vital for clarifying the reasoning behind AI decisions, evaluating them through human judgments is complex and resource-intensive due to subjectivity and the need for fine-grained ratings. This study explores the alignment between ChatGPT and human assessments across multiple scales (i.e., binary, ternary, and 7-Likert scale). We sample 300 data instances from three NLE datasets and collect 900 human annotations for both informativeness and clarity scores as the text quality measurement. We further conduct paired comparison experiments under different ranges of subjectivity scores, where the baseline comes from 8,346 human annotations. Our results show that ChatGPT aligns better with humans in more coarse-grained scales. Also, paired comparisons and dynamic prompting (i.e., providing semantically similar examples in the prompt) improve the alignment. This research advances our understanding of large language models' capabilities to assess the text explanation quality in different configurations for responsible AI development.

Understanding News Thumbnail Representativeness by Counterfactual Text-Guided Contrastive Language-Image Pretraining

This paper delves into the critical challenge of understanding the representativeness of news thumbnail images, which often serve as the first visual engagement for readers when an article is disseminated on social media. We focus on whether a news image represents the main subject discussed in the news text. To serve the challenge, we introduce NewsTT, a manually annotated dataset of news thumbnail image and text pairs. We found that pretrained vision and language models, such as CLIP and BLIP-2, struggle with this task. Since news subjects frequently involve named entities or proper nouns, a pretrained model could not have the ability to match its visual and textual appearances. To fill the gap, we propose CFT-CLIP, a counterfactual text-guided contrastive language-image pretraining framework. We hypothesize that learning to contrast news text with its counterfactual, of which named entities are replaced, can enhance the cross-modal matching ability in the target task. Evaluation experiments using NewsTT show that CFT-CLIP outperforms the pretrained models, such as CLIP and BLIP-2. Our code and data will be made accessible to the public after the paper is accepted.

K-HATERS: A Hate Speech Detection Corpus in Korean with Target-Specific Ratings

Numerous datasets have been proposed to combat the spread of online hate. Despite these efforts, a majority of these resources are English-centric, primarily focusing on overt forms of hate. This research gap calls for developing high-quality corpora in diverse languages that also encapsulate more subtle hate expressions. This study introduces K-HATERS, a new corpus for hate speech detection in Korean, comprising approximately 192K news comments with target-specific offensiveness ratings. This resource is the largest offensive language corpus in Korean and is the first to offer target-specific ratings on a three-point Likert scale, enabling the detection of hate expressions in Korean across varying degrees of offensiveness. We conduct experiments showing the effectiveness of the proposed corpus, including a comparison with existing datasets. Additionally, to address potential noise and bias in human annotations, we explore a novel idea of adopting the Cognitive Reflection Test, which is widely used in social science for assessing an individual's cognitive ability, as a proxy of labeling quality. Findings indicate that annotations from individuals with the lowest test scores tend to yield detection models that make biased predictions toward specific target groups and are less accurate. This study contributes to the NLP research on hate speech detection and resource construction. The code and dataset can be accessed at https://github.com/ssu-humane/K-HATERS.

Detecting Contextomized Quotes in News Headlines by Contrastive Learning

Quotes are critical for establishing credibility in news articles. A direct quote enclosed in quotation marks has a strong visual appeal and is a sign of a reliable citation. Unfortunately, this journalistic practice is not strictly followed, and a quote in the headline is often "contextomized." Such a quote uses words out of context in a way that alters the speaker's intention so that there is no semantically matching quote in the body text. We present QuoteCSE, a contrastive learning framework that represents the embedding of news quotes based on domain-driven positive and negative samples to identify such an editorial strategy. The dataset and code are available at https://github.com/ssu-humane/contextomized-quote-contrastive.

How does fake news use a thumbnail? CLIP-based Multimodal Detection on the Unrepresentative News Image

This study investigates how fake news uses a thumbnail for a news article with a focus on whether a news article's thumbnail represents the news content correctly. A news article shared with an irrelevant thumbnail can mislead readers into having a wrong impression of the issue, especially in social media environments where users are less likely to click the link and consume the entire content. We propose to capture the degree of semantic incongruity in the multimodal relation by using the pretrained CLIP representation. From a source-level analysis, we found that fake news employs a more incongruous image to the main content than general news. Going further, we attempted to detect news articles with image-text incongruity. Evaluation experiments suggest that CLIP-based methods can successfully detect news articles in which the thumbnail is semantically irrelevant to news text. This study contributes to the research by providing a novel view on tackling online fake news and misinformation. Code and datasets are available at https://github.com/ssu-humane/fake-news-thumbnail.

Who Is Missing? Characterizing the Participation of Different Demographic Groups in a Korean Nationwide Daily Conversation Corpus

A conversation corpus is essential to build interactive AI applications. However, the demographic information of the participants in such corpora is largely underexplored mainly due to the lack of individual data in many corpora. In this work, we analyze a Korean nationwide daily conversation corpus constructed by the National Institute of Korean Language (NIKL) to characterize the participation of different demographic (age and sex) groups in the corpus.

Who Blames or Endorses Whom? Entity-to-Entity Directed Sentiment Extraction in News Text

Understanding who blames or supports whom in news text is a critical research question in computational social science. Traditional methods and datasets for sentiment analysis are, however, not suitable for the domain of political text as they do not consider the direction of sentiments expressed between entities. In this paper, we propose a novel NLP task of identifying directed sentiment relationship between political entities from a given news document, which we call directed sentiment extraction. From a million-scale news corpus, we construct a dataset of news sentences where sentiment relations of political entities are manually annotated. We present a simple but effective approach for utilizing a pretrained transformer, which infers the target class by predicting multiple question-answering tasks and combining the outcomes. We demonstrate the utility of our proposed method for social science research questions by analyzing positive and negative opinions between political entities in two major events: 2016 U.S. presidential election and COVID-19. The newly proposed problem, data, and method will facilitate future studies on interdisciplinary NLP methods and applications.

Understanding Effects of Editing Tweets for News Sharing by Media Accounts through a Causal Inference Framework

To reach a broader audience and optimize traffic toward news articles, media outlets commonly run social media accounts and share their content with a short text summary. Despite its importance of writing a compelling message in sharing articles, research community does not own a sufficient level of understanding of what kinds of editing strategies are effective in promoting audience engagement. In this study, we aim to fill the gap by analyzing the current practices of media outlets using a data-driven approach. We first build a parallel corpus of original news articles and their corresponding tweets that were shared by eight media outlets. Then, we explore how those media edited tweets against original headlines, and the effects would be. To estimate the effects of editing news headlines for social media sharing in audience engagement, we present a systematic analysis that incorporates a causal inference technique with deep learning; using propensity score matching, it allows for estimating potential (dis-)advantages of an editing style compared to counterfactual cases where a similar news article is shared with a different style. According to the analyses of various editing styles, we report common and differing effects of the styles across the outlets. To understand the effects of various editing styles, media outlets could apply our easy-to-use tool by themselves.