We invite two categories of submissions:

  1. Full papers: Up to 7 pages (excluding references) using the IJCAI template describing novel work, with additional pages allowed in the appendix.
  2. Hot-off-the-press: describing previously published, peer-reviewed work that appeared on or after 1 January 2025. Authors are required to submit the PDF of their published work and an additional paragraph explaining the relevance of the work to the workshop.

LLM usage policy: submissions are allowed to use LLMs for assistance in writing. However, the authors are responsible for all content, including checking for plagiarised content and correctness of the submission. Hallucinated references and facts will lead to desk rejection of the submission.

Themes and Scope

The workshop is open for submissions at the intersection of AutoAI and foundation models, including, but not limited to the following topics:

  • Foundation models for AutoAI

  • AutoAI for foundation models

  • Natural language interfaces for AutoAI systems

  • In-context learning for AutoAI and meta-algorithmics

  • Zero-shot and few-shot AutoAI

  • Benchmarking and reproducibility of AutoAI methods in the foundation-model era

  • Green and resource-efficient AutoAI with foundation models

  • Safety, robustness, and reliability of foundation models via AutoAI

  • AutoAI for fine-tuning and adaptation

  • AutoAI for pretraining, post-training, and distillation

  • Model selection, routing, and orchestration using AutoAI

  • AutoAI for agentic and tool-using foundation models

  • Automated scaling analysis and performance extrapolation for foundation models

  • Interpretability and explanation of algorithms, models, and performance

(Note: In the topics above, we use ‘Foundation Models’ to broadly refer to Large Language Models (LLMs) as well as foundation models for other modalities.)

Review Policy

  • The review process is single blind.
  • Full papers will go through the following review policy:
    • All submissions will undergo review by at least three reviewers per paper.
    • Submissions will be evaluated based on technical soundness, readability and novelty.
    • Acceptance or rejection will be based on the reviewers’ assessments.
  • Hot-off-the-press papers will be evaluated by their relevance to the workshop and the quality of the venue where they got published.

Submission Instructions

Papers should use the IJCAI authors kit and be submitted via the → OpenReview submission website.

Important Dates

  • Submission deadline: 4 May 2026 (AoE)
  • Notification: 11 June 2026 (AoE)
  • Workshop dates: 15, 16 or 17 August 2026 (TBD)