Saved in:
Bibliographic Details
Main Authors: Lerman, Sam, Bi, Jing
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2506.21367
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913914172735488
author Lerman, Sam
Bi, Jing
author_facet Lerman, Sam
Bi, Jing
contents rQdia regularizes Q-value distributions with augmented images in pixel-based deep reinforcement learning. With a simple auxiliary loss, that equalizes these distributions via MSE, rQdia boosts DrQ and SAC on 9/12 and 10/12 tasks respectively in the MuJoCo Continuous Control Suite from pixels, and Data-Efficient Rainbow on 18/26 Atari Arcade environments. Gains are measured in both sample efficiency and longer-term training. Moreover, the addition of rQdia finally propels model-free continuous control from pixels over the state encoding baseline.
format Preprint
id arxiv_https___arxiv_org_abs_2506_21367
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle rQdia: Regularizing Q-Value Distributions With Image Augmentation
Lerman, Sam
Bi, Jing
Machine Learning
Artificial Intelligence
rQdia regularizes Q-value distributions with augmented images in pixel-based deep reinforcement learning. With a simple auxiliary loss, that equalizes these distributions via MSE, rQdia boosts DrQ and SAC on 9/12 and 10/12 tasks respectively in the MuJoCo Continuous Control Suite from pixels, and Data-Efficient Rainbow on 18/26 Atari Arcade environments. Gains are measured in both sample efficiency and longer-term training. Moreover, the addition of rQdia finally propels model-free continuous control from pixels over the state encoding baseline.
title rQdia: Regularizing Q-Value Distributions With Image Augmentation
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2506.21367