A Distributional View on Multi-Objective Policy Optimization

被引:0
|
作者
Abdolmaleki, Abbas [1 ]
Huang, Sandy H. [1 ]
Hasenclever, Leonard [1 ]
Neunert, Michael [1 ]
Song, H. Francis [1 ]
Zambelli, Martina [1 ]
Martins, Murilo F. [1 ]
Heess, Nicolas [1 ]
Hadsell, Raia [1 ]
Riedmiller, Martin [1 ]
机构
[1] DeepMind, London, England
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 problems require trading off multiple competing objectives. However, these objectives are often in different units and/or scales, which can make it challenging for practitioners to express numerical preferences over objectives in their native units. In this paper we propose a novel algorithm for multi-objective reinforcement learning that enables setting desired preferences for objectives in a scale-invariant way. We propose to learn an action distribution for each objective, and we use supervised learning to fit a parametric policy to a combination of these distributions. We demonstrate the effectiveness of our approach on challenging high-dimensional real and simulated robotics tasks, and show that setting different preferences in our framework allows us to trace out the space of nondominated solutions.
引用
收藏
页数:12
相关论文
共 50 条
  • [1] Distributional learning in multi-objective optimization of recommender systems
    Candelieri A.
    Ponti A.
    Giordani I.
    Bosio A.
    Archetti F.
    Journal of Ambient Intelligence and Humanized Computing, 2023, 14 (08) : 10849 - 10865
  • [2] Evolutionary Multi-objective Optimization for Multi-view Clustering
    Jiang, Bo
    Qiu, Feiyue
    Yang, Shipin
    Wang, Liping
    2016 IEEE CONGRESS ON EVOLUTIONARY COMPUTATION (CEC), 2016, : 3308 - 3315
  • [3] Distributional Multi-Objective Decision Making
    Ropke, Willem
    Hayes, Conor F.
    Mannion, Patrick
    Howley, Enda
    Nowe, Ann
    Roijers, Diederik M.
    PROCEEDINGS OF THE THIRTY-SECOND INTERNATIONAL JOINT CONFERENCE ON ARTIFICIAL INTELLIGENCE, IJCAI 2023, 2023, : 5711 - 5719
  • [4] Evolutionary multi-objective optimization: A historical view of the field
    Coello Coello, Carlos A.
    IEEE COMPUTATIONAL INTELLIGENCE MAGAZINE, 2006, 1 (01) : 28 - 36
  • [5] Pessimistic Off-Policy Multi-Objective Optimization
    Alizadeh, Shima
    Bhargava, Aniruddha
    Gopalswamy, Karthick
    Jain, Lalit
    Kveton, Branislav
    Liu, Ge
    INTERNATIONAL CONFERENCE ON ARTIFICIAL INTELLIGENCE AND STATISTICS, VOL 238, 2024, 238
  • [6] A DISTRIBUTIONAL ROBUST MINIMAX REGRET OPTIMIZATION APPROACH FOR MULTI-OBJECTIVE PORTFOLIO PROBLEMS
    Ou, Xiaoqing
    Song, Dan
    Zhang, Tao
    Chen, Jiawei
    JOURNAL OF NONLINEAR AND CONVEX ANALYSIS, 2024, 25 (08) : 1881 - 1897
  • [7] Multi-objective stable matching and distributional constraints
    Gharote, Mangesh
    Phuke, Nitin
    Patil, Rahul
    Lodha, Sachin
    SOFT COMPUTING, 2019, 23 (09) : 2995 - 3011
  • [8] Multi-objective stable matching and distributional constraints
    Mangesh Gharote
    Nitin Phuke
    Rahul Patil
    Sachin Lodha
    Soft Computing, 2019, 23 : 2995 - 3011
  • [9] A new hybrid memetic multi-objective optimization algorithm for multi-objective optimization
    Luo, Jianping
    Yang, Yun
    Liu, Qiqi
    Li, Xia
    Chen, Minrong
    Gao, Kaizhou
    INFORMATION SCIENCES, 2018, 448 : 164 - 186
  • [10] Multi-Objective Optimization of a Port-of-Entry Inspection Policy
    Young, Christina M.
    Li, Mingyu
    Zhu, Yada
    Xie, Minge
    Elsayed, Elsayed A.
    Asamov, Tsvetan
    14TH ISSAT INTERNATIONAL CONFERENCE ON RELIABILITY AND QUALITY IN DESIGN, PROCEEDINGS, 2008, : 172 - +