5 papers
AllDayNav: Lifelong Navigation via Real-World Reinforcement Learning
Hang Yin, Yinan Liang, Jiazhao Zhang +4
Lifelong embodied navigation in dynamic environments requires robots to form persistent scene understanding from fragmentary observations, which remains difficult for existing meth…
Reducing Variance in Meta-Learning via Laplace Approximation for Regression Tasks
Alfredo Reichlin, Gustaf Tegnér, Miguel Vasco +3
Given a finite set of sample points, meta-learning algorithms aim to learn an optimal adaptation strategy for new, unseen tasks. Often, this data can be ambiguous as it might belon…
Interactive Perception for Deformable Object Manipulation
Zehang Weng, Peng Zhou, Hang Yin +4
Interactive perception enables robots to manipulate the environment and objects to bring them into states that benefit the perception process. Deformable objects pose challenges to…
Goal-Conditioned Reinforcement Learning from Sub-Optimal Data on Metric Spaces
Alfredo Reichlin, Miguel Vasco, Hang Yin +1
We study the problem of learning optimal behavior from sub-optimal datasets for goal-conditioned offline reinforcement learning under sparse rewards, invertible actions and determi…
Learning Geometric Representations of Objects via Interaction
Alfredo Reichlin, Giovanni Luca Marchetti, Hang Yin +2
We address the problem of learning representations from observations of a scene involving an agent and an external object the agent interacts with. To this end, we propose a repres…