analiza aiqbench
SUSD addresses a key limitation in unsupervised skill discovery by factorizing state spaces to capture dynamic behaviors across multiple environmental entities, outperforming existing MI-based and distance-maximizing methods in complex settings. By allocating distinct skill variables to independent components, the framework enables fine-grained control and facilitates efficient training of downstream hierarchical reinforcement learning tasks. This approach offers a more robust foundation for autonomous agents to learn diverse, task-relevant skills without relying on extrinsic rewards.
SUSD: Structured Unsupervised Skill Discovery through State Factorization
arXiv:2602.01619v2 Announce Type: replace Abstract: 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. Distance-Maximizing Skill Discovery (DSD) promotes more dynamic skills by leveraging state-space distances, yet still fall short in encouraging comprehensive skill sets that engage all controllable factors or entities in the environment. In this work, we introduce SUSD, a novel framework that harnesses the compositional structure of environments by factorizing the state space into independent components (e.g., objects or controllable entities). SUSD allocates distinct skill variables to different factors, enabling more fine-grained control on the skill discovery process. A dynamic model also tracks learning across factors, adaptively steering the agent's focus toward underexplored factors. This structured approach not only promotes the discovery of richer and more diverse skills, but also yields a factorized skill representation that enables fine-grained and disentangled control over individual entities which facilitates efficient training of compositional downstream tasks via Hierarchical Reinforcement Learning (HRL). Our experimental results across three environments, with factors ranging from 1 to 10, demonstrate that our method can discover diverse and complex skills without supervision, significantly outperforming existing unsupervised skill discovery methods in factorized and complex environments. Code is publicly available at: https://github.com/hadi-hosseini/SUSD.
Abstract: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. Distance-Maximizing Skill Discovery (DSD) promotes more dynamic skills by leveraging state-space distances, yet still fall short in encouraging comprehensive skill sets that engage all controllable factors or entities in the environment. In this work, we introduce SUSD, a novel framework that harnesses the compositional structure of environments by factorizing the state space into independent components (e.g., objects or controllable entities). SUSD allocates distinct skill variables to different factors, enabling more fine-grained control on the skill discovery process. A dynamic model also tracks learning across factors, adaptively steering the agent's focus toward underexplored factors. This structured approach not only promotes the discovery of richer and more diverse skills, but also yields a factorized skill representation that enables fine-grained and disentangled control over individual entities which facilitates efficient training of compositional downstream tasks via Hierarchical Reinforcement Learning (HRL). Our experimental results across three environments, with factors ranging from 1 to 10, demonstrate that our method can discover diverse and complex skills without supervision, significantly outperforming existing unsupervised skill discovery methods in factorized and complex environments. Code is publicly available at: this https URL.
Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
arXiv:2602.01619 [cs.LG]
(or arXiv:2602.01619v2 [cs.LG] for this version)
https://doi.org/10.48550/arXiv.2602.01619 Focus to learn more arXiv-issued DOI via DataCite
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 335 cuvinte
Subiecte conexe
Supervised Deep Multimodal Matrix Factorization for Interpretable Brain Network Analysis
arXiv · 28 sept. 2026
Common Cause, Not Cross-Attention: Blocking Visual Shortcuts in Audio-Video Generation
arXiv · 22 sept. 2026
When Identical Rows Disagree: From Benchmark Identifiability to Replication-Robust Anomaly Detection
arXiv · 26 sept. 2026
From Directions to Regions: Decomposing Activations in Language Models via Local Geometry
arXiv · 30 sept. 2026