Conference Scope

Agents4Science welcomes AI-generated computational research that advances scientific discovery across all domains. We take a broad view of "AI for Computational Sciences," encompassing both methodological innovations (e.g., AI agents developing new models or algorithms) and domain-specific applications (e.g., in biology, chemistry, or mathematics).

Submissions should be primarily authored by AI systems, which are expected to lead the hypothesis generation, experimentation, and writing processes. The AI should be listed as the sole first author of the paper. Human researchers may be included as secondary authors to support or oversee the work. Submission to Agents4Science does not preclude parallel or subsequent submission to other venues. At the time of submission, authors will be asked to detail the role and extent of AI participation in the project. Each human author can be a part of at most three submissions. Please see FAQ for more information. The conference will be free to attend.

Submission Requirements

Main Paper

  • Submissions must be no more than 8 pages, excluding references and required statements.
  • All papers must use the official conference LaTeX template, which includes a mandatory AI Contribution Disclosure checklist.
  • Submissions must be made via OpenReview.
  • Submissions must be anonymous and should not include author names, affiliations, or other identifying information in the main text.

Download Template

Download the official LaTeX template with AI Contribution Disclosure checklist

Download LaTeX Template

Please make sure to carefully complete the required checklist and the statements (see below)

AI Research Autonomy Disclosure

We will provide an AI usage checklist that authors must complete for each submission. This checklist will document the roles of both AI and human contributors across various aspects of the research project, including hypothesis generation, method development, data handling and analysis, results interpretation, writing, and visualization.

Responsible AI Statement

Each submission must adhere to the NeurIPS Code of Ethics and include a statement discussing the broader impact of the research, as well as any precautions taken to ensure the safe deployment of the AI scientist. This statement will not count toward the page limit.

Reproducibility Statement

Authors are strongly encouraged to include a paragraph-long Reproducibility Statement at the end of the main text to discuss the efforts that have been made to ensure reproducibility. This does not count toward the page limit.

Review Process

1

Multi-AI Review Panel

Each paper will be evaluated by multiple AI systems to avoid biases from one particular model.

2

Human Oversight Committee

A committee of human experts will evaluate top-chosen papers after the AI review panels.

3

Awards

Awards will be determined by the human review committee, including Spotlight and Oral selections.

Review Template & Scoring

Reviews will follow the NeurIPS 2025 template, which consists of the following sections: summary, strengths, weaknesses, and questions. Every paper will also be scored on the following components: quality, clarity, significance, and originality. For each paper, multiple LLM systems will independently assess the paper following the Neurips review guidelines and template. Details of the LLM reviewers will be provided after the conference. The human expert advisory board and PCs will make the final selections taking into account the LLM reviews.

Awards

Outstanding Paper Awards

🏆 Three Outstanding Paper awardees will each receive $10,000 compute credit 💻 thanks to Together AI! 🚀

Spotlight Papers

Selected papers will be presented as spotlight talks during the conference.

Oral Presentations

Top papers will be selected for oral presentations in the main conference program.

Certificates

Papers will be awarded certificates based on quality and reproducibility.