Efficient model-free exploration in low-rank mdps
A major challenge in reinforcement learning is to develop practical, sample-efficient
algorithms for exploration in high-dimensional domains where generalization and function …
algorithms for exploration in high-dimensional domains where generalization and function …
Representation learning with multi-step inverse kinematics: An efficient and optimal approach to rich-observation rl
Z Mhammedi, DJ Foster… - … Conference on Machine …, 2023 - proceedings.mlr.press
We study the design of sample-efficient algorithms for reinforcement learning in the
presence of rich, high-dimensional observations, formalized via the Block MDP problem …
presence of rich, high-dimensional observations, formalized via the Block MDP problem …
Model-based rl with optimistic posterior sampling: Structural conditions and sample complexity
We propose a general framework to design posterior sampling methods for model-based
RL. We show that the proposed algorithms can be analyzed by reducing regret to Hellinger …
RL. We show that the proposed algorithms can be analyzed by reducing regret to Hellinger …
Inverse dynamics pretraining learns good representations for multitask imitation
D Brandfonbrener, O Nachum… - Advances in Neural …, 2024 - proceedings.neurips.cc
In recent years, domains such as natural language processing and image recognition have
popularized the paradigm of using large datasets to pretrain representations that can be …
popularized the paradigm of using large datasets to pretrain representations that can be …
Online control of unknown time-varying dynamical systems
E Minasyan, P Gradu… - Advances in Neural …, 2021 - proceedings.neurips.cc
We study online control of time-varying linear systems with unknown dynamics in the
nonstochastic control model. At a high level, we demonstrate that this setting is\emph …
nonstochastic control model. At a high level, we demonstrate that this setting is\emph …
Smoothed online learning for prediction in piecewise affine systems
A Block, M Simchowitz… - Advances in Neural …, 2024 - proceedings.neurips.cc
The problem of piecewise affine (PWA) regression and planning is of foundational
importance to the study of online learning, control, and robotics, where it provides a …
importance to the study of online learning, control, and robotics, where it provides a …
Non-asymptotic and accurate learning of nonlinear dynamical systems
We consider the problem of learning a nonlinear dynamical system governed by a nonlinear
state equation ht+ 1= ϕ (ht, ut; θ)+ wt. Here θ is the unknown system dynamics, ht is the …
state equation ht+ 1= ϕ (ht, ut; θ)+ wt. Here θ is the unknown system dynamics, ht is the …
Learning mixtures of linear dynamical systems
Y Chen, HV Poor - International conference on machine …, 2022 - proceedings.mlr.press
We study the problem of learning a mixture of multiple linear dynamical systems (LDSs) from
unlabeled short sample trajectories, each generated by one of the LDS models. Despite the …
unlabeled short sample trajectories, each generated by one of the LDS models. Despite the …
Non-linear reinforcement learning in large action spaces: Structural conditions and sample-efficiency of posterior sampling
Abstract Provably sample-efficient Reinforcement Learning (RL) with rich observations and
function approximation has witnessed tremendous recent progress, particularly when the …
function approximation has witnessed tremendous recent progress, particularly when the …
A reinforcement learning look at risk-sensitive linear quadratic gaussian control
In this paper, we propose a robust reinforcement learning method for a class of linear
discrete-time systems to handle model mismatches that may be induced by sim-to-real gap …
discrete-time systems to handle model mismatches that may be induced by sim-to-real gap …