Call for full papers

The 50th Annual ACM SIGIR Conference on Research and Development in Information Retrieval

The annual SIGIR conference is the major international forum for the presentation of new research results, and the demonstration of new systems and techniques, in the broad field of information retrieval (IR). The 50th ACM SIGIR conference will be run as an in-person conference in July 2027 in San Jose, California.

For the full paper call, we welcome high-impact original papers with contributions related to any aspect of information retrieval and access, including theories, foundations, algorithms, evaluation, analysis, and applications. Please note that the CFPs for other paper tracks, as well as workshops, tutorials, doctoral consortium, industry day, and other SIGIR 2027 venues will be released separately.

Important Dates (AoE) — PROPOSED

Full paper abstract submission: January 14, 2027
Full paper submission: January 21, 2027
Full paper notification: April 5, 2027
Camera-ready submission: April 30, 2027
Author registration: April 30, 2027

Full paper submission is not possible without a previously submitted abstract.

New in 2027

This year's Full Paper CFP introduces several changes from prior years. Here is a summary:

Agentic IR is now a standalone topic area. Reflecting the rapid growth of agent-based information access, from tool-augmented retrieval to autonomous multi-step search, we have elevated Agentic IR into its own top-level area, separate from Conversational IR.
New topic area: Information Access through Generative AI. We introduce a dedicated area for research on how generative models are changing the way people access information, covering attribution, grounding, retrieval integration in compound AI systems, and the interplay between retrieval-based and generation-based access.
Expanded FATE area to include AI Safety in IR. The Fairness, Accountability, Transparency, Ethics area now explicitly covers adversarial robustness, misinformation in AI-mediated search, guardrails for agentic systems, and alignment of search with user intent.
Consolidated "Foundation Models for IR" area. We merged the former "Machine Learning for IR" and "Natural Language Processing for IR" areas into a single area reflecting the convergence of these fields around foundation models. The emphasis is on models that advance retrieval and access, not on general model research evaluated on IR benchmarks.
Updated evaluation area. We added explicit encouragement for evaluation frameworks addressing agentic and multi-step IR, groundedness and hallucination measurement, and cost-quality tradeoffs.
Chair for Integrity and Robustness. We are introducing a dedicated chair role to support the review process with automated checks on research integrity and methodological robustness. This complements and does not replace expert reviewer judgment.
Satellite presentation option. Presenting authors who are unable to obtain a visa, or who have serious safety or other significant concerns about attending in person, may apply to present at a satellite location instead. Requests are reviewed through an approval process. Please check Satellite Presentation Policy for eligibility and how to apply.

Relevant Areas

Relevant topic areas include (but are not limited to):

Search and Ranking

Research on core IR algorithmic topics, such as:

Queries and query analysis.
Web search.
Retrieval models and ranking.
Theoretical models and foundations of information retrieval and access.

System, Efficiency and Scalability

Research on search system aspects that relate to the efficiency of the system and/or its scalability, such as:

Efficient and scalable indexing, crawling, compression, search, and more.
Energy efficiency and green computing for IR.
Search engine architecture, distributed search, metasearch, peer-to-peer search, search in the cloud, edge IR.
Inference cost optimization for neural retrieval and generative components.

Recommender Systems

Research focusing on recommender systems, rich content representations and content analysis for recommendation, such as:

Filtering and recommendation.
Cross-domain recommendation, socially-aware and context-aware recommender systems, multi-stakeholder recommendations.
LLM-based and conversational recommendation.
Recommendation in agentic workflows and compound AI systems.
Other theoretical models and foundations of recommender systems.

Foundation Models for IR (formerly "Machine Learning for IR" and "NLP for IR")

Research that uses or adapts foundation models to advance information retrieval and access. We emphasize contributions where the retrieval or access problem drives the research, rather than general model papers evaluated on IR benchmarks. Topics include:

Deep learning and representation learning for IR.
Fine-tuning, distillation, and adaptation of foundation models for retrieval.
Generative retrieval models (e.g., generative document IDs, direct answer generation with attribution).
Retrieval Augmented Generation (RAG) and retrieval-augmented reasoning.
Reasoning-augmented retrieval (e.g., chain-of-thought for search, inference-time compute scaling).
Reinforcement learning and learning from interactions for IR.
Click models and implicit feedback.
Question answering.

Agentic IR (new — expanded from "Conversational or Agentic IR")

Research on IR systems that autonomously plan, use tools, and execute multi-step information tasks with minimal user intervention:

Tool-augmented and API-augmented retrieval.
Multi-step search planning and execution.
Autonomous web agents for information tasks.
Agentic RAG, including orchestrating retrieval, reasoning, and action.
Intelligent personal assistants and agents.

Conversational IR

Research on interactive, user-driven information access through multi-turn dialog:

End-to-end conversational IR models and optimization.
Session-based search and recommendation, user engagement.
Conversational question answering, dialog systems, spoken language interfaces.
Mixed-initiative interaction, when the system and user jointly steer the search process.

Information Access through Generative AI (new)

Research on how generative models are changing the way people find and access information, such as:

Attribution, grounding, and source provenance in generative search.
Retrieval as the grounding layer in generative AI systems.
User information-seeking behavior in the era of answer engines (e.g., zero-click search, changing query patterns).
Retrieval quality, ranking robustness, and credibility assessment in corpora with AI-generated content.
Comparison and complementarity of retrieval-based and generation-based information access.

Humans and Interfaces

Research into user-centric aspects of IR including user interfaces, behavior modeling, privacy, interactive systems, such as:

User studies, qualitative, and quantitative.
User interfaces and visualization.
Social and collaborative search.
User modeling.
Human-AI collaboration in IR and sensemaking.

Datasets, Benchmarks, and Evaluation for IR

Research that focuses on the measurement and evaluation of IR systems, such as:

Benchmarks and test collections.
User-centered evaluation.
New methods for building data sets.
Online evaluation.
Evaluation frameworks for agentic and multi-step IR.
Measuring groundedness, hallucination, and factual accuracy in generative retrieval.
Evaluation of agentic IR systems (correctness, safety, cost, latency).
Simulation for evaluation.
Metrics and evaluation methodology.

Fairness, Accountability, Transparency, Ethics, Safety, and Explainability (FATES) in IR (expanded)

Research on FATES aspects and bias in search systems and related applications:

Fairness, accountability, transparency and explainability.
Ethics, economics, and politics.
Adversarial robustness of retrieval-augmented and agentic systems.
Misinformation, manipulation, and content integrity in AI-mediated search.
Guardrails, alignment, and safety in agentic IR.

Multi Modal IR

Theoretical, algorithmic or novel practical solutions addressing problems across the domain of multimedia and IR, such as:

Multimedia search and retrieval (e.g., image search, video search, speech and audio search, music search).
Maps and spatial search.
Multi-modal embeddings and cross-modal retrieval.

Domain-Specific IR Applications

Research focusing on domain-specific IR challenges, such as:

Local and mobile search.
Social search.
Search in structured data.
Education.
Legal.
Health.
Scientific literature search and discovery.
Other applications and domains.

Full Paper Submission Guidelines

See this brief checklist to strengthen an IR paper, for authors and reviewers.

Full research papers must describe original work that has not been previously published (except on pre-print servers, see below for details), not accepted for publication elsewhere, and not simultaneously submitted or currently under review in another journal or conference (including the other tracks of SIGIR 2027). Please note that concurrent submissions are a violation of the ACM Policy on Authorship. The SIGIR 2027 program chairs will refer any such violations to the Ethics & Plagiarism Committee of the ACM Publications Board.

Submissions of full research papers must be in English, in PDF format, and be at most 9 pages (including figures, tables, proofs, appendixes, acknowledgments, and any content except references) in length, with unrestricted space for references, in the current ACM two-column conference format. Suitable LaTeX, Word, and Overleaf templates are available from the ACM Website (use "sigconf" proceedings template for LaTeX and the Interim Template for Word). ACM's CCS concepts and keywords are required for review.

For LaTeX, the following should be used:

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

Submissions must be anonymous and should be submitted electronically via OpenReview.

Accepted papers will be allowed to add one extra page (10 pages plus references) to the final version of the paper, to give the authors the ability to incorporate reviewer feedback into the final version.

SIGIR 2027 Common Submission Requirements

Submitting to this track indicates your agreement to and compliance with the common ACM and SIGIR 2027 Policies and Requirements. These policies and requirements include (but are not limited to) authorship, ethics, the use of AI, ACM author open access page charges, and at least one author registering for and attending the conference.

Authorship Policy

Authors should carefully read the ACM's authorship policy before submitting to SIGIR 2027.

By submitting your article to an ACM Publication, you are hereby acknowledging that you and your co-authors are subject to all ACM Publications Policies, including ACM's Publications Policy on Research Involving Human Participants and Subjects. Alleged violations of this policy or any ACM Publications Policy will be investigated by ACM and may result in a full retraction of your paper, in addition to other potential penalties, as per ACM Publications Policy.

To support the identification of reviewers with conflicts of interest, the full author list must be specified at abstract submission time. No changes to authorship, under any circumstances, will be permitted after the abstract submission deadline or for the camera-ready submission. Please make sure that you have listed authors correctly at abstract submission time.

OpenReview Registration and Profile

All authors must be registered in OpenReview before the abstract submission deadline. Signing up for a profile can take up to two weeks, especially if you use a public email address (like Gmail). To expedite the process, use an institutional email address from your university or company, as these are often approved automatically.

All authors must complete their profile in OpenReview, and be responsive to any emails sent by the review system. Each author must fully complete their profile information. Regarding the "Personal Links", each author with a DBLP profile and at least 10 of their publications listed in that profile must provide the corresponding DBLP URL and click the "Add DBLP Papers to Profile" button. Authors with fewer of their publications in DBLP are encouraged to enter the corresponding DBLP URL, provided that it only contains their own publications; otherwise, the URL of an ORCID page with all the author's previous publications should be entered, and the "Add ORCID Papers to Profile" button should be clicked. We also encourage entering the other links (Homepage, Google Scholar, ORCID, etc.) but they are not mandatory.

Use of AI

All submissions must comply with the ACM policy on the use of Artificial Intelligence.

Desk Rejection Policy

In conjunction with the Common Submission Requirements linked above, note that for Full Paper submissions any of the following may (and likely will) result in desk rejection:

An incomplete profile in OpenReview as explained above; in particular, each author with a DBLP profile and at least 10 of their publications listed in that profile must provide the corresponding DBLP URL and click the "Add DBLP Papers to Profile" button.
Change of authors after the abstract submission deadline.
Figures, tables, proofs, appendixes, acknowledgements, or any other content except for the references after page 9 of the submission.
Formatting that is not in line with the guidelines provided above.
Clear lack of topical fit for SIGIR.
Obvious anonymity violations. For example, authors or authors' institutional affiliations clearly named in the submission.
Links to source code repositories that directly reveal the identities or institutional affiliations of the authors. SIGIR 2027 explicitly welcomes links to source code and artifacts as part of the submission, but please use anonymous repositories, such as https://anonymous.4open.science.
Content that has been determined to have been copied from other sources. Cases of plagiarism (including self-plagiarism) will be reported to the Ethics & Plagiarism Committee of the ACM Publications Board.
Any form of academic fraud or dishonesty, according to ACM's policies on academic dishonesty.

Author Reviewing Expectations

Peer-reviewed conferences such as SIGIR 2027 rely on a large and knowledgeable pool of reviewers. For each submission to SIGIR 2027, we ask one of the authors to agree to serve as a reviewer. In the case of authors submitting multiple papers, we ask for a different author to be nominated for each submission. We will do our best to assign reviews according to the reviewing author's expertise.

Review Process

Each manuscript will be reviewed by at least three PC members and a senior PC member. All reviewing will be double-anonymous. The acceptance decisions will take into account manuscript novelty, technical depth, elegance, practical or theoretical impact, and presentation.

The peer review process follows the guidelines laid out in the ACM Peer Review Policy.

Reviewers are expected to disclose any conflicts of interest they may have (see the ACM Conflict of Interest Policy), treat submissions as confidential, carefully read their assigned submissions, and provide fair, thoughtful and respectful reviews.

Here are examples of "highly irresponsible" reviews:

Reviews that are not submitted in a timely fashion.
Reviews that lack content are generic or irrelevant to the submission.
Reviews that insult the authors.
Reviews that are generated by a large language model; see https://www.acm.org/publications/policies/peer-review-faq.

If a review is flagged as "highly irresponsible", it will undergo an oversight process managed by the Program Chairs. Any reviewer whose review is deemed to be "highly irresponsible" will be reported to the SIGIR Executive Committee, and at the discretion of the Executive Committee may be banned from submitting work to future instances of conferences sponsored by SIGIR.

Submission System

Papers must be submitted as PDF files via OpenReview (link to the submission site to come).

Program Chairs

program-chairs@sigir2027.org

Mounia Lalmas, Spotify, UK
Qiaozhu Mei, University of Michigan, USA
Mark Sanderson, RMIT University, Australia