SyncAI.news, a Varaisys broadcasting
StarWM: Self-Supervised Trained Attention Routing for Robust World Models
ZZ

Zeqiang Zhang, Fabian Wurzberger, Maximilian Otte, Daniel Schmid, Sebastian Gottwald, Arne Peter Raulf, Daniel Alexander Braun

· 1 min read

ResearcharXiv cs.CV

StarWM: Self-Supervised Trained Attention Routing for Robust World Models

arXiv:2609.30667v1 Announce Type: new Abstract: A robust world model must strike the balance between faithfully capturing environmental dynamics and abstracting away from irrelevant content. While reconstruction-based world models ensure faithful supervision, they misallocate representational capacity by pixel area rather than dynamics relevance for visual tasks, which can cause task-irrelevant content to dominate the learned representation. Alternatively, reconstruction-free methods avoid this bias but risk discarding possibly relevant information. We propose StarWM, which uses a cross-attention module trained on self-supervised dynamics to decide where reconstruction applies. A dual-stream decoder then restricts reconstruction to the attended regions, with stop-gradient barriers preventing interference between the two objectives. These components allows reconstruction to supervise the visual content of attended regions without contaminating the latent with non-predictive information. On DeepMind Control with dynamic video backgrounds, default (reward-free) StarWM achieves the strongest performance under random-frame distractors and substantially outperforms reconstruction-based baselines under sequential video. In addition, its reward-augmented variant matches or exceeds reconstruction-free methods on sequential video, achieving the highest overall return across all distractor regimes. Mechanistic probing confirms StarWM preserves state attributes with near-perfect fidelity through long-horizon imagination while systematically discarding distractors.

Original source

This story was published by arXiv cs.CV and written by Zeqiang Zhang, Fabian Wurzberger, Maximilian Otte, Daniel Schmid, Sebastian Gottwald, Arne Peter Raulf, Daniel Alexander Braun. SyncAI.news shows a preview; the complete article is on the publisher's site.

Read the full story on arxiv.org

Similar News