CLEF tracks presentations

AI is increasingly central to understanding and managing biodiversity and ecosystems. Since 2011, the LifeCLEF lab has provided large-scale benchmarks that stimulate progress in multimodal species recognition, ecological prediction, and knowledge extraction. The 2026 edition expands this scope with five complementary challenges spanning visual, acoustic, and textual data: (i) \textbf{AnimalCLEF}: discovery and re-identification of individual animals, (ii) \textbf{BirdCLEF+}: multi-taxonomic species recognition in complex soundscapes, (iii) \textbf{MarineCLEF}: detection of marine species in underwater imagery under positive-unlabeled constraints, (iv) \textbf{PestCLEF}: extraction of information on plant pests from heterogeneous textual sources, (v) \textbf{PlantCLEF}: multi-species plant identification in quadrat images. Together, these challenges address critical dimensions of biodiversity science and ecosystem management, while fostering collaboration between AI researchers, ecologists, and practitioners. This paper provides an overview of the LifeCLEF 2026 lab and its tasks, outlining their motivation, data, and evaluation methodology to guide participants and inform the wider research community.

Since its inception in 2003, the various ImageCLEF challenges have provided large and complex datasets targeting a wide array of subjects in medicine, argumentation, reasoning, content recommendation, data generation, and question answering. In its 24th edition, ImageCLEF will have five main tasks: (i) a Medical task, which aims to promote the synergy between four medical challenges: Caption, involving concept detection and caption prediction in radiology images, Synthetic Medical Image Generation in the GANs task, Visual Question Answering for improving the diagnosis and classification of real medical gastrointestinal images, and multimodal dermatology response generation and a new MEDIQA-CORE challenge focusing on predicting or correcting tumor type labels and identifying and summarizing major and minor differences between pairs of radiology reports; (ii) the ToPicto task, involving text to pictogram translation and prediction, (iii) the Multimodal Reasoning task on visual, multi-language, interdisciplinary question answering, and 2 new tasks: (iv) AI4Agriculture, involving predicting agricultural potential before planing and crop type identification and (v) the Deepfake detection and generation task. In its last edition, 56 teams finished our challenges, continuing to show the impact in the community.

Decision-making and opinion-forming are everyday tasks that involve weighing pro and con arguments. The goal of Touch¬_ is to foster the development of support-technologies for decision-making and opinion-forming and to improve our understanding of these processes. This seventh edition of the lab features four shared tasks, of which the first three are new in 2026: (1) Fallacy Detection (new task), in which participants determine whether an argument follows a valid argument pattern or a fallacy; (2) Causality Extraction (new task), in which participants extract pro- and concausal claims from natural language text; (3) Generalizability of Argument Identification in Context (new task), in which participants predict whether or not sentences would be annotated as an argument under different annotation guidelines; and (4) Advertisement in Retrieval-Augmented Generation (2nd edition), in which participants detect and block advertisements in generated text. In this paper, we briefly describe the planned setup for the seventh lab edition at CLEF 2026 and summarize the results of the 2025 edition.

Quantum Computing (QC) is a research field that has been in the limelight in recent years. In fact, this new paradigm has the potential to revolutionize the way we currently solve problems by leveraging quantum-mechanical phenomena, which allow quantum computers to solve specific problems more efficiently than traditional computers. As quantum computers are starting to become more available, our objective is to investigate the application of QC within the Information Retrieval (IR) and Recommender Systems (RS) fields. In fact, IR and RS systems perform computationally intensive operations on extensive datasets, and using QC in their pipeline could be useful to improve their efficiency and, in some cases, effectiveness. Thus, in this work, we present the third edition of the QuantumCLEF lab, the first lab that allows participants to use real quantum computers for solving IR and RS tasks. The lab is composed of three main tasks that aim at discovering and evaluating Quantum Annealing (QA) approaches compared to their traditional counterpart while also establishing collaborations among researchers from different fields to harness their knowledge and skills to solve the considered challenges and promote the usage of QA. Moreover, if quantum resources are available, we plan to introduce a gate-based task for more-experienced researchers.

The main objective of the CheckThat! lab is to advance the development of innovative technologies combating disinformation and manipulation efforts in online communication across a multitude of languages and platforms. Focusing on core tasks of the verification pipeline in earlier years, namely check-worthiness, evidence retrieval, and verification, the CheckThat! lab broadened its focus during the past three editions and included auxiliary challenges linked to the verification process. In the 2026 edition of the lab, the verification pipeline is placed at the center again, further expanding the core tasks of the pipeline: Task 1 is on source retrieval for scientific web claims (a follow-up of the CheckThat! 2025 edition), Task 2 focuses on fact-checking numerical and temporal claims, adding a reasoning component to the 2025 edition of the task, and Task 3 expands the verification pipeline by a task on the generation of full-fact-checking articles. These tasks represent challeng ing classification and retrieval problems as well as generation challenges at the document and at the span level, including multilingual settings.

HIPE-2026 is a CLEF evaluation lab that focuses on person--place relation extraction from noisy, multilingual historical texts. Building on the previous HIPE campaigns in 2020 and 2022, this lab introduces a new task that targets semantic relations between persons and places. Specifically, systems are asked to classify person--place relations across two relation types: 1. At---``Has the person been at this place?'' and 2. isAt---``Is the person at the given place?'' (relative to a document's publication date). The task requires an understanding of the text and the ability to reason about temporal and geographical information, and is designed to be tackled by both generative AI systems (large language models) and more traditional classification approaches. The lab promotes both accuracy and efficiency by offering dual evaluation profiles and testing generalisation capabilities with a surprise test set. HIPE-2026 aims to support downstream applications in knowledge graph construction, historical biography reconstruction, and spatial analysis in digital humanities. We present the motivation, task design, datasets, and evaluation framework for the lab.

Large generative language models (LLMs) are a foundational technology for both research and information service development, and are used as a tool for solving challenging text analysis problems. The ELOQUENT lab develops shared tasks for experimenting with the resilience of large language models applied to practical tasks. In 2026, the tasks will include the Voight Kampff task for exploring how well one can distinguish human-authored text from machine-generated text, the Robustness and Consistency task to explore how well a model adapts to the culture expressed by a language, and the Quiz task, to experiment with generating and scoring test quiz material for use in real educational situations.

The paper gives a brief overview of the four shared tasks organized in the PAN 2026 lab on digital text forensics and stylometry to be hosted at CLEF 2026. The goal of the PAN lab is to advance the state-of-the-art in text forensics and stylometry through an objective evaluation of new and established methods on new benchmark datasets. Our five tasks in 2026 will be: (1) Voight-Kampff Generative AI Detection, particularly in mixed and obfuscated authorship scenarios, (2) Text Watermarking, a new task that aims to find new and benchmark the robustness of existing text watermarking schemes, (3) Multi-author Writing Style Analysis, a continued task that aims to find positions of authorship change, (4) Generative Plagiarism Detection, a continued task that targets source retrieval and text alignment between generated text and source documents, (5) Reasoning Trajectory Detection, a new task that deals with the source detection and safety detection of LLM-generated or human-written reasoning trajectories. As in previous editions, PAN invites software submissions as easy-to-reproduce docker containers; more than 1,100 of softwares have been submitted since PAN~2012, with all recent evaluations running on the TIRA experimentation platform.

Since its foundation in 2017, the eRisk CLEF Lab has pioneered research in early risk detection on the Internet, focusing on mental health challenges such as depression, anorexia, and pathological gambling. Over the years, participants have contributed to the development of detection models and exploited the datasets we constructed to advance this critical area. In 2026, which marks the tenth edition of the lab, we continue this trajectory with three tasks that emphasize conversational and contextual modeling as well as symptom-oriented retrieval. The first task, Conversational Depression Detection, introduces the challenge of identifying depression through interactions with fine-tuned Large Language Models (LLMs) personas, which this year will be openly accessible via Hugging Face. The second task, Contextualised Early Detection of Depression, focuses on user-level classification by analyzing full conversational contexts, with participants engaging iteratively in natural interactions. Finally, the third task, ADHD Symptom Sentence Ranking, expands our scope beyond depression by requiring systems to rank sentences according to their relevance to the symptoms defined in the Adult ADHD Self-Report Scale (ASRS-v1.1). This paper outlines the progress of the lab to date, introduces the three tasks of eRisk 2026, and discusses our innovative plans for fostering research on mental health challenges.

Over the past thirteen years, the large-scale biomedical semantic indexing and question-answering challenge (BioASQ) has consistently driven the advancement of methods and tools that enhance access to the rapidly growing body of scientific resources in the biomedical domain. BioASQ offers a unique common testbed where research teams worldwide can evaluate and compare innovative approaches for accessing biomedical knowledge by organizing shared tasks on an annual basis and providing respective benchmark datasets that represent the real information needs of biomedical experts. The fourteenth version of BioASQ will be held as an evaluation Lab in the context of CLEF2026, providing six tasks: (i) Task b on biomedical semantic question answering. (ii) Task Synergy on question answering for developing biomedical topics. (iii) Task MultiClinSum-2 on multilingual clinical summarization. (iv) Task BioNNE-R on nested relation extraction in Russian and English. (v) Task ELCardioCC on clinical coding in cardiology. (vi) Task GutBrainIE on gut-brain interplay information extraction. Through its six shared tasks, the 14th edition of BioASQ challenges the research community to develop methods that go beyond the current state of the art, fostering innovative approaches that enable efficient and precise access to biomedical knowledge while pushing the research frontier forward.