Reinforcement Learning
Enabling Adaptive Agent Training in Open-Ended Simulators by Targeting Diversity
The wider application of end-to-end learning methods to embodied decision-making domains remains bottlenecked by their reliance on a superabundance of training data representative of the target domain.
Disentangled Unsupervised Skill Discovery for Efficient Hierarchical Reinforcement Learning
However, existing unsupervised skill discovery methods often learn entangled skills where one skill variable simultaneously influences many entities in the environment, making downstream skill chaining extremely challenging.
Amortizing intractable inference in diffusion models for vision, language, and control
Code is available at this link .
Local Linearity: the Key for No-regret Reinforcement Learning in Continuous MDPs
Existing solutions either work under very specific assumptions or achieve bounds that are vacuous in some regimes.
On Divergence Measures for Training GFlowNets
In reinforcement learning (RL), a recurring goal is to find a diverse set of high-valued state-action trajectories according to a reward function.
