NeurIPS 2026 workshop calls for continual world models that keep learning after deployment

jonLorraine9 · x · 2026-08-19

The Continual World Models workshop at NeurIPS 2026 (Sydney, December) is calling for submissions on world models that keep learning after deployment — "a world model should change when the world does."

It runs two non-archival tracks: Main for substantive technical work (up to 8 pages) and Idea for 2-4 page papers on early directions, positions, new formulations, negative results, and lessons from failed experiments. Submissions close Aug 29 AoE via OpenReview.

Related event: NeurIPS 2026 Workshop on Continual World Models Calls for Papers(3 posts)→

Original post →

More from Research

Research channel →