Workshop on AI for the Scholarly Web (NSLP-SDP 2027)

10/11 May 2027 – Co-located with The Web Conference (WebConf) 2027 Dublin, Ireland

https://nfdi4ds.github.io/nslp-sdp-2027/

In addition to the regular workshop paper track, NSLP-SDP 2027 features four shared tasks:

* ClimateCheck 2027: Cross-Lingual Scientific Fact-Checking of Social Media Claims about Climate Change * SciVQA 2027: Scientific Visual Question Answering * IE4ML 2027: Information Extraction for Machine Learning * READNER 2027: Named Entity Recognition in GitHub READMEs

IMPORTANT DATES (FOR BOTH WORKSHOP PAPERS AND SHARED TASK PAPERS)

* Submission deadline: 11 January 2027 * Notifications: 08 February 2027 * Camera-ready: 16 February 2027

The scholarly Web has become the primary infrastructure through which scientific knowledge is published, discovered, linked, assessed, and reused. Making this ecosystem machine-actionable is simultaneously a document-understanding problem and a knowledge-acquisition-and-use problem. Scientific knowledge is communicated through long, structurally complex and increasingly multi-modal documents containing specialized terminology, citations, tables, figures, equations, metadata, code, datasets, and other research artefacts. Researchers and research infrastructures need machine-actionable representations of scientific claims, methods, evidence, entities, relations, and provenance in order to search, verify, synthesize, and reuse the rapidly growing body of scientific literature.

The 1st Joint Workshop on AI for the Scholarly Web (NSLP-SDP 2027) unifies two established and complementary workshop series. Natural Scientific Language Processing (NSLP) has emphasized scientific-language processing, information extraction, research/scientific knowledge graphs (RKGs/SKGs), research artefact identification/linking, metadata, multilinguality, scientific IR, and research infrastructures. Scholarly Document Processing (SDP) has built a strong community around document understanding, multimodal scholarly content, discourse/argumentation, citation analysis, summarization, QA, bibliometrics/scientometrics, peer review, reproducibility, and research integrity. The joint NSLP-SDP workshop attempts to connect the full pipeline: scholarly Web content => structured and grounded scientific knowledge => Web-scale search, synthesis and trustworthy research agents.

SDP has been held at EMNLP 2020https://ornlcda.github.io/SDProc/, NAACL 2021https://sdproc.org/2021/, COLING 2022https://sdproc.org/2022/, ACL 2024https://sdproc.org/2024/, and ACL 2025https://sdproc.org/2025/. NSLP was held at ESWC 2024https://nfdi4ds.github.io/nslp2024/, ESWC 2025https://nfdi4ds.github.io/nslp2025/, and LREC 2026https://nfdi4ds.github.io/nslp2026/. The 2027 edition differs fundamentally from previous editions by merging the NSLP and SDP communities, moving beyond an NLP-centred venue toward the broader scholarly Web.

TOPICS OF INTEREST INCLUDE, BUT ARE NOT LIMITED TO

* LLMs, multimodal models, representation learning, domain adaptation, and scientific-language resources; * Scholarly document understanding: structure & layout, figures, tables, equations, and multimodal grounding; * IE, NER, linking & coreference, relations & events, extraction of datasets, software, methods, tasks, results, claims; * Scholarly and research KGs, linked research objects, semantic annotation, Web-scale metadata integration, provenance, interoperability, machine-actionable scientific knowledge; * Web-scale scholarly search and recommendation, retrieval-augmented AI, scientific QA and synthesis, literature discovery, and autonomous or human-in-the-loop research agents; * Citation & context analysis, discourse & argumentation, novelty & contribution modeling, and claim & evidence linking; * Peer review and writing & reading support, research integrity & reproducibility, generated-content detection, bibliometrics & scientometrics, and science-of-science; * Factuality, uncertainty, robustness, explainability, reproducible evaluation, and responsible & human-centered deployment; * Scholarly Web content, FAIR and open Web resources, digital libraries, repository ecosystems, interoperable research infrastructures, cross-platform scholarly data integration.

SUBMISSION GUIDELINES

The NSLP-SDP 2027 workshop invites methodological, resource, evaluation, system, and position papers. We especially encourage submissions from first-time authors, students, researchers from underrepresented groups and regions, and work on non-English scholarly communication.

* We will not accept work that is under review or has already been published in or accepted for publication in a journal, another conference, or another workshop. * The workshop invites anonymous submissions of regular long papers (5-8 pages without references and appendix); short papers as well as position papers (3-4 pages without references and appendix) presenting, for example, negative results, in-progress projects, or demos. * Authors are permitted to include an optional appendix of up to 2 pages. However, reviewers will not be mandated to review the appendix and all papers must be self-contained. * Submissions are expected to discuss relevant limitations and risks of the work, for example fabricated claims and citations, over-reliance on automated review, assessment, or writing, bias towards English and well-resourced disciplines, privacy and confidentiality, copyright, licensing and provenance, or compute and environmental costs. * Reviewing will be double-blind. Each submission will receive three reviews. * Submissions must be in PDF, formatted in the WebConf 2027 style, i.e., the usual ACM proceedings two-column page style. In LaTeX, please use: \documentclass[sigconf, review]{acmart}. * The proceedings of this workshop will be published as part of the dedicated WebConf 2027 Companion Proceedings volume. * NSLP-SDP 2027 is an in-person workshop. At least one author per contribution must register for the workshop and present in person. * All submissions are done via OpenReview – the link will be provided soon.

SHARED TASKS

IE4ML 2027: Shared Task on Information Extraction for Machine Learning

Reproducing ML research requires structured access to the models, datasets, methods, and relations described in scientific publications. IE4ML builds on GSAP-ERE, an extension of GSAP-NER, and turns its schema into a community benchmark for producing machine-actionable descriptions suitable for scholarly knowledge graphs and Web-scale discovery. The task uses 10 entities and 18 relation types and comprises two subtasks: 1. IE4ML Entity Extraction (span extraction and classification); 2. IE4ML Relation Extraction (given gold entity spans). Training uses 80 publicly released GSAP-ERE publications; validation and test sets each contain 20 newly annotated papers released during the evaluation phases.

Shared task co-organisers: Wolfgang Otto, Kanishka Silva, Gan Lu, Sharmila Upadhyaya, Stefan Dietze

READNER 2027: Named Entity Recognition in GitHub READMEs

READNER 2027 is a fine-grained Named Entity Recognition (NER) shared task. It addresses research-related GitHub README files, where metadata about datasets, software, experimental code, and other research artifacts is often embedded in free-form documentation. GitHub repositories complement research papers by providing datasets, models, source code, and implementation details that help researchers explore and reproduce results. Their README files typically describe the project’s purpose, setup, and usage, and link to related research resources. READNER extracts structured metadata from README documents for research resource discovery and reuse. Participants develop systems that take a GitHub README file as input and identify entity mentions, their boundaries, and their types.

Shared task co-organisers: Genet Asefa Gesese, Zongxiong Chen, Mary Ann Tan, Sonja Schimmler

ClimateCheck 2027: Cross-Lingual Scientific Fact-Checking of Social Media Claims about Climate Change

Social media claims about climate change appear across languages, yet scientific fact-checking resources, including previous ClimateCheck iterations (run at SDP 2025 and NSLP 2026), have largely paired claims and evidence only within English. ClimateCheck 2027 introduces a cross-lingual setting: given a claim in German or Arabic, systems must retrieve supporting or refuting evidence from English-language environmental-science abstracts. The languages span different language families and scripts, enabling evaluation of cross-lingual retrieval and evidence grounding. Claims will be derived from previous English ClimateCheck datasets and validated by native speakers. Subtask 1 – Abstract retrieval: given a claim in (Levantine) Arabic and/or German, and corpus of publications, retrieve the top 10 most relevant abstracts. Subtask 2 – Claim verification: given the claim-abstract pair from subtask 1, classify whether the abstract supports, refutes, or does not have enough evidence about the claim.

Shared task co-organisers: Raia Abu Ahmad, Georg Rehm

SciVQA 2027: Scientific Visual Question Answering

Scientific papers communicate information through unstructured text as well as (semi-)structured figures and tables. Jointly reasoning over both modalities benefits downstream applications such as visual question answering (VQA). SciVQA 2027 builds on the insights from SciVQA 2025https://sdproc.org/2025/scivqa.html (run at SDP 2025), shifting the focus toward evaluating the ability of multimodal LLMs to reason over combined modalities (figures, tables, text). SciVQA 2027 introduces a new set of Computer Science and Computational Linguistics papers, with a total of 1,500 manually annotated QA pairs. The shared task features two subtasks: Evidence retrieval – Given a paper and a question, retrieve candidate evidence (paragraphs and/or images) containing information needed to derive the answer. Answer generation – Given a paper and a question, generate an answer.

Shared task co-organisers: Ekaterina Borisova, Georg Rehm

WORKSHOP ORGANISERS

* Georg Rehm (DFKI & Humboldt-Universität zu Berlin, Germany) * Philipp Mayr (GESIS Leibniz Institute for the Social Sciences, Germany) * Tirthankar Ghosal (Oak Ridge National Laboratory, USA) * Stefan Dietze (GESIS Leibniz Institute for the Social Sciences & Heinrich-Heine-University Düsseldorf, Germany) * Yufang Hou (IT:U Interdisciplinary Transformation University, Austria & IBM Research) * Aakanksha Naik (Allen Institute for AI, USA) * Dan Li (Elsevier, The Netherlands) * Sonja Schimmler (Fraunhofer FOKUS & Technical University of Berlin, Germany) * Anita de Waard (Elsevier, USA)

CONTACT

* Main contact: Georg Rehm <georg.rehm@dfki.demailto:georg.rehm@dfki.de> * https://nfdi4ds.github.io/nslp-sdp-2027/