Non-archival contributed papers
Up to 6 pages in NeurIPS format (references not counted); accepted papers are presented as posters or spotlight orals and published as non-archival reports allowing future archival submissions.
Machine Learning/CORE Unranked
WRL@NeurIPS 2026
8th Robot Learning Workshop: Is Physical AI Going Zero-Shot?
The 8th Robot Learning Workshop at NeurIPS 2026 explores whether Physical AI is transitioning to zero-shot learning through foundation models, agentic reasoning, and embodied generalization. It invites research on scaling laws, multi-modal models, real-world deployment, and evaluation benchmarks to examine the future of robotics beyond task-specific fine-tuning.
Paper fit
Contribution paths
A strong submission should clearly identify its contribution and evaluate it appropriately.
Research areas in scope
01
Topics
Zero-shot, few-shot, and continual adaptation for robot learningVision-language-action models and generalist robot policiesWorld models and generative simulatorsCross-embodiment and cross-environment generalizationLarge-scale robot data, simulation, and data curationEvaluation protocols and benchmarks for zero-shot Physical AISafety, uncertainty, and failure analysis in deployed robotsReal-world deployment studies and negative results
Policies worth checking twice
- Submissions must be anonymized for double-blind review (no author names, affiliations, or identifying information)
- Up to 6 pages in NeurIPS format (references do not count toward the page limit)
- Submissions consisting solely of previously published conference papers are discouraged without substantial new discussion or insight
- Accepted papers are non-archival and may be submitted to archival conferences or journals
Official sources
Compiled from the official call for papers. The organizers’ pages remain authoritative.
Last verified September 9, 2026