FlexiWorld builds on Joint Embedding Predictive Architectures (JEPAs) that predict future latent states without pixel reconstruction.
Flexible action chunks boost world model planning for distant goals
FlexiWorld uses variable-length action chunks and mixed-span goal supervision to achieve 89.29% mean success across four benchmarks, outperforming fixed-chunk baselines.
Big Tech
Shidu Ren · Qilin Gu · Zhenghao Ni · Junhan Sun · Jiaqi Wang · Damien Scieur · +1 more
University of Toronto · Zhejiang University · Tencent Jarvis Lab · Mila & Universite de Montreal · Samsung SAIL
Research Digest··3 min read
The authors introduce FlexiWorld, a JEPA-based world model that jointly learns latent prediction and goal-conditioned action generation from action chunks of varying length and goal spans.
Why this paper
From Samsung SAIL and 5 others
In one line
FlexiWorld achieves 89.29% mean success via variable-length action chunks and mixed-span goal supervision.
What we could check
- ·No code link found
- ·No weights link found
- ·No dataset link found
- ·No compute details found
- ✓Limitations stated by the authors
- ✓Reports numbers on named benchmarks
Observed from the paper text and links we have. Absence here means we did not find it, not that it does not exist.
§