Meta-Reinforcement Learning by Tracking Task Non-stationarity

被引:0
|
作者
Poiani, Riccardo [1 ]
Tirinzoni, Andrea [2 ]
Restelli, Marcello [1 ]
机构
[1] Politecn Milan, Milan, Italy
[2] Inria Lille, Lille, France
关键词
D O I
暂无
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Many real-world domains are subject to a structured non-stationarity which affects the agent's goals and the environmental dynamics. Meta-reinforcement learning (RL) has been shown successful for training agents that quickly adapt to related tasks. However, most of the existing meta-RL algorithms for non-stationary domains either make strong assumptions on the task generation process or require sampling from it at training time. In this paper, we propose a novel algorithm (TRIO) that optimizes for the future by explicitly tracking the task evolution through time. At training time, TRIO learns a variational module to quickly identify latent parameters from experience samples. This module is learned jointly with an optimal exploration policy that takes task uncertainty into account. At test time, TRIO tracks the evolution of the latent parameters online, hence reducing the uncertainty over future tasks and obtaining fast adaptation through the meta-learned policy. Unlike most existing methods, TRIO does not assume Markovian task-evolution processes, it does not require information about the non-stationarity at training time, and it captures complex changes undergoing in the environment. We evaluate our algorithm on different simulated problems and show it outperforms competitive baselines.
引用
收藏
页码:2899 / 2905
页数:7
相关论文
共 50 条
  • [1] Meta-Reinforcement Learning via Exploratory Task Clustering
    Chu, Zhendong
    Cai, Renqin
    Wang, Hongning
    THIRTY-EIGHTH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE, VOL 38 NO 10, 2024, : 11633 - 11641
  • [2] Non-stationarity in multiagent reinforcement learning in electricity market simulation
    Renshaw-Whitman, Charles
    Zobernig, Viktor
    Cremer, Jochen L.
    de Vries, Laurens
    ELECTRIC POWER SYSTEMS RESEARCH, 2024, 235
  • [3] Deep Reinforcement Learning amidst Continual Structured Non-Stationarity
    Xie, Annie
    Harrison, James
    Finn, Chelsea
    INTERNATIONAL CONFERENCE ON MACHINE LEARNING, VOL 139, 2021, 139
  • [4] REACTIVE EXPLORATION TO COPE WITH NON-STATIONARITY IN LIFELONG REINFORCEMENT LEARNING
    Steinparz, Christian
    Schmied, Thomas
    Paischer, Fabian
    Dinu, Marius-Constantin
    Patil, Vihang
    Bitto-Nemling, Angela
    Eghbal-zadeh, Hamid
    Hochreiter, Sepp
    CONFERENCE ON LIFELONG LEARNING AGENTS, VOL 199, 2022, 199
  • [5] Decoupling Meta-Reinforcement Learning with Gaussian Task Contexts and Skills
    He, Hongcai
    Zhu, Anjie
    Liang, Shuang
    Chen, Feiyu
    Shao, Jie
    THIRTY-EIGHTH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE, VOL 38 NO 11, 2024, : 12358 - 12366
  • [6] Model-Free Non-Stationarity Detection and Adaptation in Reinforcement Learning
    Canonaco, Giuseppe
    Restelli, Marcello
    Roveri, Manuel
    ECAI 2020: 24TH EUROPEAN CONFERENCE ON ARTIFICIAL INTELLIGENCE, 2020, 325 : 1047 - 1054
  • [7] Harnessing Meta-Reinforcement Learning for Enhanced Tracking in Geofencing Systems
    Famili, Alireza
    Sun, Shihua
    Atalay, Tolga
    Stavrou, Angelos
    IEEE Open Journal of the Communications Society, 2025, 6 : 944 - 960
  • [8] Information-theoretic Task Selection for Meta-Reinforcement Learning
    Gutierrez, Ricardo Luna
    Leonetti, Matteo
    ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 33, NEURIPS 2020, 2020, 33
  • [9] Generalizable Task Representation Learning for Offline Meta-Reinforcement Learning with Data Limitations
    Zhou, Renzhe
    Gao, Chen-Xiao
    Zhang, Zongzhang
    Yu, Yang
    THIRTY-EIGHTH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE, VOL 38 NO 15, 2024, : 17132 - 17140
  • [10] Meta-Reinforcement Learning Based on Self-Supervised Task Representation Learning
    Wang, Mingyang
    Bing, Zhenshan
    Yao, Xiangtong
    Wang, Shuai
    Kai, Huang
    Su, Hang
    Yang, Chenguang
    Knoll, Alois
    THIRTY-SEVENTH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE, VOL 37 NO 8, 2023, : 10157 - 10165