Call for Resource Papers

The 50th ACM SIGIR conference takes place in San Jose, California, from July 18 to 24, 2027. The conference is held in person, with a satellite location in Tianjin, China, at which accepted papers can also be presented under the conditions of the Satellite Presentation Policy. The Resources and Reproducibility tracks continue to be offered as two separate tracks, as introduced at SIGIR 2026. This is the call for papers for the Resources Track.

Submissions will be peer-reviewed, and accepted papers will be published in the main conference proceedings. Resource papers will be reviewed single-anonymous so that reviewers can examine shared resources that cannot be anonymized. All submissions must include author names and affiliations.

Papers must not exceed 6 pages, with unlimited space for references. Shorter submissions are welcome. Reviewers will assess the length of a paper relative to its contribution.

Key Dates (AoE)

A paper can be submitted only if its abstract, title and complete author list were submitted by the abstract deadline. Placeholder abstracts will be rejected after the abstract deadline, and the author list cannot be changed afterwards.

Scope of the Track

Resources may support any aspect of information retrieval and access within the scope of the SIGIR 2027 call for full papers. Resources include, but are not restricted to:

• Test collections and relevance judgments for information retrieval and access tasks.
• Datasets and corpora for training, adapting and evaluating models for retrieval and information access, including foundation models.
• Documentation of task designs and evaluation protocols.
• Software tools and services for information retrieval and access, including the evaluation and analysis of retrieval systems.
• Models, indexes and embeddings built for IR tasks.
• Environments and simulators for information retrieval and access tasks.

Resources for the two new areas of SIGIR 2027, agentic IR and information access through generative AI, are particularly welcome. The review criteria are the same for all topics.

Resources Built with Generative AI

Resources created partly or entirely with generative AI are eligible. This includes synthetic documents and queries, model-generated relevance labels, code written with coding agents, and LLM-based judges.

The paper must describe how the generated parts were produced and how their quality was checked, in enough detail for a reader to understand and reproduce the process. Reviewers will assess this description under Utility.

Authors remain responsible for the resource and the paper, including every generated part. The use of generative AI in writing the paper must be disclosed, as the ACM policy on authorship requires.

Making the Resource Available

The resource must be available to reviewers from submission until notification, through a link given in the paper, so that reviewers can confirm that the resource matches what is described. The link must work without login, or the access instructions must be self-contained in the submission.

For very large resources, a representative sample or a lightweight example should be provided where feasible. The license or terms of use must be stated in the paper.

Persistent identifiers and archival hosting are encouraged, as is a short statement of how the resource will remain available and how future versions will be managed.

Reviewing

Resource papers will be reviewed single-anonymous. Reviewers remain anonymous and must assess the resource without regard to the authors' reputation or institution. Reviews that do otherwise will be replaced. Reviewers will consider the following questions.

Novelty. What is new about the resource? Is it an incremental advance or a more substantial one? Which existing resources are closest to it, and what does it add?

Availability. Is the resource available to the reviewer at the time of review? Are there discrepancies between what is described and what is available? Are the license and terms of use open enough for most academic and industry researchers? If the resource contains data collected from people, were the appropriate human-subjects procedures followed?

Utility. Is the resource well documented, and what level of expertise is needed to use it? Are there tutorials or examples, and do they reflect actual use? If the resource is data, are tools provided for loading it, and are the provenance stages (source, preprocessing, cleaning, aggregation) documented? If parts of the resource were generated or labeled by a model, is the generation process documented, and were the outputs checked?

Predicted impact. What IR research activity does the resource enable? Does it advance an established research area or open a new one? Is it expected to remain useful over time, or will it need curation or updates, and if so, is that planned? How large is the anticipated research user community, and is it likely to grow or shrink?

There will be no rebuttal phase.

Submitting

Submissions must be in English and in PDF, in the ACM two-column conference format.

Suitable LaTeX, Word and Overleaf templates are available from the ACM Website. Use the sigconf proceedings template for LaTeX and the ACM Interim Template for Word. For LaTeX, use:

\documentclass[sigconf,natbib=true,review]{acmart}

Length is at most 6 pages, including figures, tables, appendices and acknowledgements, with unlimited space for references. Shorter submissions are welcome. Reviewers will assess the length of a paper relative to its contribution. CCS concepts and keywords are not required for review but are required for publication, and their inclusion in the initial submission is recommended. Accepted papers may add one extra page (7 pages plus references) in the camera-ready version to address reviewer feedback.

Papers must be submitted through EasyChair, in the "SIGIR 2027 Resource Papers" track: https://easychair.org/my/conference?conf=sigir2027. All submissions must include author names and affiliations. Each submission must nominate, in the submission form, one of its authors as a reviewer for the track. The same person cannot be the nominee for more than one submission.

Submitting to this track indicates agreement to and compliance with the common ACM and SIGIR 2027 policies, including those on authorship, ethics, the use of AI, plagiarism, and ACM author open access page charges. Preprints are permitted under the SIGIR 2027 preprint policy. A paper may not be under review in another venue or another SIGIR 2027 track. A resource paper and a paper in another track may not describe the construction of the same resource. Submissions that exceed the page limit, do not follow the format, add authors after the abstract deadline, violate the preprint policy, or provide no working access to the resource will be rejected without review. At least one author must register for the conference by April 30, 2027 and present the paper at the conference.

Resources Track Chairs

resources@sigir2027.org

• Saber Zerhoudi, University of Passau, Germany
• Jiaxin Mao, Renmin University of China, China