AI RESEARCH
SUSD: Structured Unsupervised Skill Discovery through State Factorization
arXiv CS.AI
•
ArXi:2602.01619v2 Announce Type: replace-cross Unsupervised Skill Discovery (USD) aims to autonomously learn a diverse set of skills without relying on extrinsic rewards. One of the most common USD approaches is to maximize the Mutual Information (MI) between skill latent variables and states. However, MI-based methods tend to favor simple, static skills due to their invariance properties, limiting the discovery of dynamic, task-relevant behaviors.