paper-with-me

홈 › Papers

Exponential Family Model-Based Reinforcement Learning via Score Matching

2021-12-28 · Gene Li, Junbo Li, Anmol Kabra, Nathan Srebro, Zhaoran Wang, Zhuoran Yang

We propose an optimistic model-based algorithm, dubbed SMRL, for finite-horizon episodic reinforcement learning (RL) when the transition model is specified by exponential family distributions with $d$ parameters and the reward is bounded and known. SMRL uses score matching, an unnormalized density estimation technique that enables efficient estimation of the model parameter by ridge regression. Under standard regularity assumptions, SMRL achieves $\tilde O(d\sqrt{H^3T})$ online regret, where $H$ is the length of each episode and $T$ is the total number of interactions (ignoring polynomial dependence on structural scale parameters).

📄 PDF Abstract BibTeX arXiv:2112.14195

Code (1)

anmolkabra/score-matching-rl 공식 구현 pytorch

Tasks

Density EstimationModel-based Reinforcement Learningregressionreinforcement-learningReinforcement LearningReinforcement Learning (RL)

Similar Papers 제목 키워드 기반

Robust Score Matching

2025-01-09 · Richard Schwank, Andrew McCormack, Mathias Drton

Proposed in Hyv\"arinen (2005), score matching is a parameter estimation procedure that does not require computation of distributional normalizing constants. In this work we utilize the geometric median of means to devel…

parameter estimation

Kernel Exponential Family Estimation via Doubly Dual Embedding

2018-11-06 · Bo Dai, Hanjun Dai, Arthur Gretton, Le Song 외

We investigate penalized maximum log-likelihood estimation for exponential family distributions whose natural parameter resides in a reproducing kernel Hilbert space. Key to our approach is a novel technique, doubly dual…

Provable benefits of score matching

2023-06-03 · NeurIPS 2023 11

Score matching is an alternative to maximum likelihood (ML) for estimating a probability distribution parametrized up to a constant of proportionality. By fitting the ''score'' of the distribution, it sidesteps the need …

Finite Sample Bounds for Learning with Score Matching

2026-05-13 · Devin Smedira, Abhijith Jayakumar, Sidhant Misra, Marc Vuffray 외 arxiv

Learning of continuous exponential family distributions with unbounded support remains an important area of research for both theory and applications in high-dimensional statistics. In recent years, score matching has be…

High-Dimensional Differential Parameter Inference in Exponential Family using Time Score Matching

2024-10-14 · Daniel J. Williams, Leyang Wang, Qizhen Ying, Song Liu 외

This paper addresses differential inference in time-varying parametric probabilistic models, like graphical models with changing structures. Instead of estimating a high-dimensional model at each time point and estimatin…