Adversarial knowledge distillation framework that stress-tests and distills Code-LLMs into smaller, more robust and reliable models.
May 5, 2025
Unsupervised skill discovery approach using automatically generated tasks and asymmetric self-play to learn composable robotic manipulation behaviors.
Oct 7, 2024
Multi-critic actor-critic approach that combines multiple reward functions to discover safe, useful robotic manipulation skills.
Feb 1, 2024
LLM-based approach that automatically converts text-based task descriptions into reward and goal-generation functions for robotic manipulation.
Jun 19, 2023
Memory-augmented policy networks for non-Markovian control problems with external memory modules.
Jan 1, 2017
Novel adaptive sampling technique for SGD that learns optimal sampling distributions to accelerate convergence.
Jan 1, 2016