This paper presents the eighth edition of the CheckThat! lab, part of the 2025 Conference and Labs of the Evaluation Forum (CLEF). As in previous editions of CheckThat!, the lab offers tasks from the core of the verification pipeline, including check-worthiness, identifying previously fact-checked claims, supporting evidence retrieval, and claim verification as well as auxiliary tasks addressing different facets of individual steps of the pipeline: Task 1 is on identification of subjectivity (a follow-up of the CheckThat! 2024 edition), which is related to the check-worthiness task, Task 2 is on claim normalization, Task 3 addresses fact-checking numerical claims, and Task 4 focuses on scientific web discourse processing. These challenging classification and retrieval problems are offered in different mono-, multi- and crosslingual settings covering more than 20 languages. This year, CheckThat! was one of the most popular labs at CLEF-2025 in terms of team registrations: 177 teams registered, almost half of them actually participating (a total of 83 teams) and 54 submitted system description papers.

Large Language Models (LLMs) have demonstrated immense advances in a wide range of natural language tasks. However, these models are susceptible to hallucinations and errors on particularly temporal understanding tasks involving multiple entities in answers. In such tasks, they fail to associate entities with accurate time intervals, generate a complete list of entities in answers or reason about events associated with specific temporal bounds. Existing works do not extensively evaluate the abilities of the model to perform implicit and explicit temporal understanding in a list answer construction setup. To bridge this gap, we propose the Time referenced List based Question Answering or TLQA benchmark that requires structured answers in list format aligned with corresponding time periods. Our TLQA benchmark, requires both list construction and temporal understanding simultaneously, which to the best of our knowledge has not been explored in prior benchmarks. We investigate the temporal understanding and list construction capabilities of state-of-the-art generative models on TLQA in closed-book and open-domain settings. Our findings reveal significant shortcomings in current models, particularly their inability to provide complete answers and temporally align facts in a closed-book setup and the need to improve retrieval in open-domain setup, providing clear future directions for research on TLQA. The benchmark and code can be publicly accessed at https://github.com/elixir-research-group/TLQA.

Advanced relevance models, such as those that use large language models (LLMs), provide highly accurate relevance estimations. However, their computational costs make them infeasible for processing large document corpora. To address this, retrieval systems often employ a telescoping approach, where computationally efficient but less precise lexical and semantic retrievers filter potential candidates for further ranking. However, this approach heavily depends on the quality of early-stage retrieval, which can potentially exclude relevant documents early in the process. In this work, we propose a novel paradigm for re-ranking called online relevance estimation that continuously updates relevance estimates for a query throughout the ranking process. Instead of re-ranking a fixed set of top-k documents in a single step, online relevance estimation iteratively re-scores smaller subsets of the most promising documents while adjusting relevance scores for the remaining pool based on the estimations from the final model using an online bandit-based algorithm. This dynamic process mitigates the recall limitations of telescoping systems by re-prioritizing documents initially deemed less relevant by earlier stages-including those completely excluded by earlier-stage retrievers. We validate our approach on TREC benchmarks under two scenarios: hybrid retrieval and adaptive retrieval. Experimental results demonstrate that our method is sample-efficient and significantly improves recall, highlighting the effectiveness of our online relevance estimation framework for modern search systems.

The field of automated fact-checking increasingly depends on retrieving web-based evidence to determine the veracity of claims in real-world scenarios. A significant challenge in this process is not only retrieving relevant information, but also identifying evidence that can both support and refute complex claims. Traditional retrieval methods may return documents that directly address claims or lean toward supporting them, but often struggle with more complex claims requiring indirect reasoning. While some existing benchmarks and methods target retrieval for fact-checking, a comprehensive real-world open-domain benchmark has been lacking. In this paper, we present a real-world retrieval benchmark FactIR, derived from Factiverse production logs, enhanced with human annotations. We rigorously evaluate state-of-the-art retrieval models in a zero-shot setup on FactIR and offer insights for developing practical retrieval systems for fact-checking.

The CheckThat! lab aims to advance the development of innovative technologies designed to identify and to counteract online disinformation and manipulation efforts across various languages and platforms. The first five editions of the CheckThat! lab focused on the main tasks of the information verification pipeline: check-worthiness, evidence retrieval and pairing, and verification. Since the 2023 edition, the lab has broadened the focus and addressed new problems on auxiliary tasks supporting research and decision-making during the verification process. In the 2025 edition of the lab, we consider tasks at the core of the verification pipeline again as well as auxiliary tasks: Task 1 is on identification of subjectivity (a follow up of the CheckThat! 2024 edition), Task 2 is on claim normalization, Task 3 addresses fact-checking numerical claims, and Task 4 focuses on scientific web discourse processing. These tasks represent challenging classification and retrieval problems at the document and at the span level, including multilingual settings.

The advances in the digital era have led to rapid dissemination of information. This has also aggravated the spread of misinformation and disinformation. This has potentially serious consequences, such as civil unrest. While fact-checking aims to combat this, manual fact-checking is cumbersome and not scalable. While automated fact-checking approaches exist, they do not operate in real-time and do not always account for spread of misinformation through different modalities. This is particularly important as proactive fact-checking on live streams in real-time can help people be informed of false narratives and prevent catastrophic consequences that may cause civil unrest. This is particularly relevant with the rapid dissemination of information through video on social media platforms or other streams like political rallies and debates. Hence, in this work we develop a platform named LiveFC, that can aid in fact-checking live audio streams in real-time. LiveFC has a user-friendly interface that displays the claims detected along with their veracity and evidence for live streams with associated speakers for claims from respective segments.

While recent advancements in Neural Ranking Models have resulted in significant improvements over traditional statistical retrieval models, it is generally acknowledged that the use of large neural architectures and the application of complex language models in Information Retrieval (IR) have reduced the transparency of retrieval methods. Consequently, Explainability and Interpretability have emerged as important research topics in IR. Several axiomatic and post-hoc explanation methods, as well as approaches that attempt to be interpretable-by-design, have been proposed. We present ir_explain, an open-source Python library that implements a variety of well-known techniques for Explainable IR (ExIR) within a common, extensible framework. It supports the three standard categories of post-hoc explanations, namely pointwise, pairwise, and listwise explanations. The library is designed to make it easy to reproduce state-of-the-art ExIR baselines on standard test collections, as well as to explore new approaches to explaining IR models and methods. To facilitate adoption, ir_explain is well-integrated with widely-used toolkits such as Pyserini, PyTerrier (work in progress) and ir_datasets. Downstream applications of ir_explain include explaining the Retrieval-Augmented Generation (RAG) pipeline. The development version of the library is available on GitHub. We release the library as a pip package (https://pypi.org/project/ir-explain/); source code is available from https://github.com/souravsaha/ir_explain.

With the growth of misinformation on the web, automated fact checking has garnered immense interest for detecting growing misinformation and disinformation. Current systems have made significant advancements in handling synthetic claims sourced from Wikipedia, and noteworthy progress has been achieved in addressing real-world claims that are verified by fact-checking organizations as well. We compile and release QuanTemp, a diverse, multi-domain dataset focused exclusively on numerical claims, encompassing comparative, statistical, interval, and temporal aspects, with detailed metadata and an accompanying evidence collection. This addresses the challenge of verifying real-world numerical claims, which are complex and often lack precise information, a gap not filled by existing works that mainly focus on synthetic claims. We evaluate and quantify these gaps in existing solutions for the task of verifying numerical claims. We also evaluate claim decomposition based methods, numerical understanding based natural language inference (NLI) models and our best baselines achieves a macro-F1 of 58.32. This demonstrates that QuanTemp serves as a challenging evaluation set for numerical claim verification.

Online learning systems have multiple data repositories in the form of transcripts, books and questions. To enable ease of access, such systems organize the content according to a well defined taxonomy of hierarchical nature (subject - chapter -topic). The task of categorizing inputs to the hierarchical labels is usually cast as a flat multi-class classification problem. Such approaches ignore the semantic relatedness between the terms in the input and the tokens in the hierarchical labels. Alternate approaches also suffer from class imbalance when they only consider leaf level nodes as labels. To tackle the issues, we formulate the task as a dense retrieval problem to retrieve the appropriate hierarchical labels for each content. In this paper, we deal with categorizing questions and learning content. We model the hierarchical labels as a composition of their tokens and use an efficient cross-attention mechanism to fuse the information with the term representations of the content. We also adopt an adaptive in-batch hard negative sampling approach which samples better negatives as the training progresses. We demonstrate that the proposed approach TagRec++ outperforms existing state-of-the-art approaches on question and learning content datasets as measured by Recall@k. In addition, we demonstrate zero-shot capabilities of TagRec++ and preliminary analysis of it's ability to adapt to label changes.

An significant challenge in text-ranking systems is handling hard queries that form the tail end of the query distribution. Difficulty may arise due to the presence of uncommon, underspecified, or incomplete queries. In this work, we improve the ranking performance of hard or difficult queries while maintaining the performance of other queries. Firstly, we do LLM-based query enrichment for training queries using relevant documents. Next, a specialized ranker is fine-tuned only on the enriched hard queries instead of the original queries. We combine the relevance scores from the specialized ranker and the base ranker, along with a query performance score estimated for each query. Our approach departs from existing methods that usually employ a single ranker for all queries, which is biased towards easy queries, which form the majority of the query distribution. In our extensive experiments on the DL-Hard dataset, we find that a principled query performance based scoring method using base and specialized ranker offers a significant improvement of up to 48.4% on the document ranking task and up to 25% on the passage ranking task compared to the baseline performance of using original queries, even outperforming SOTA model.

As information retrieval systems continue to evolve, accurate evaluation and benchmarking of these systems become pivotal. Web search datasets, such as MS MARCO, primarily provide short keyword queries without accompanying intent or descriptions, posing a challenge in comprehending the underlying information need. This paper proposes an approach to augmenting such datasets to annotate informative query descriptions, with a focus on two prominent benchmark datasets: TREC-DL-21 and TREC-DL-22. Our methodology involves utilizing state-of-the-art LLMs to analyze and comprehend the implicit intent within individual queries from benchmark datasets. By extracting key semantic elements, we construct detailed and contextually rich descriptions for these queries. To validate the generated query descriptions, we employ crowdsourcing as a reliable means of obtaining diverse human perspectives on the accuracy and informativeness of the descriptions. This information can be used as an evaluation set for tasks such as ranking, query rewriting, or others.

Answering complex questions is a challenging task that requires question decomposition and multistep reasoning for arriving at the solution. While existing supervised and unsupervised approaches are specialized to a certain task and involve training, recently proposed prompt-based approaches offer generalizable solutions to tackle a wide variety of complex question-answering (QA) tasks. However, existing prompt-based approaches that are effective for complex QA tasks involve expensive hand annotations from experts in the form of rationales and are not generalizable to newer complex QA scenarios and tasks. We propose, icat (In-Context Ability Transfer) which induces reasoning capabilities in LLMs without any LLM fine-tuning or manual annotation of in-context samples. We transfer the ability to decompose complex questions to simpler questions or generate step-by-step rationales to LLMs, by careful selection from available data sources of related tasks. We also propose an automated uncertainty-aware exemplar selection approach for selecting examples from transfer data sources. Finally, we conduct large-scale experiments on a variety of complex QA tasks involving numerical reasoning, compositional complex QA, and heterogeneous complex QA which require decomposed reasoning. We show that ICAT convincingly outperforms existing prompt-based solutions without involving any model training, showcasing the benefits of re-using existing abilities.