Mingliang Zhang, Sichang Su, Chengyang He, Guillaume Sartoretti
{"title":"多代理强化学习中增强多任务泛化的混合训练","authors":"Mingliang Zhang, Sichang Su, Chengyang He, Guillaume Sartoretti","doi":"arxiv-2408.13567","DOIUrl":null,"url":null,"abstract":"In multi-agent reinforcement learning (MARL), achieving multi-task\ngeneralization to diverse agents and objectives presents significant\nchallenges. Existing online MARL algorithms primarily focus on single-task\nperformance, but their lack of multi-task generalization capabilities typically\nresults in substantial computational waste and limited real-life applicability.\nMeanwhile, existing offline multi-task MARL approaches are heavily dependent on\ndata quality, often resulting in poor performance on unseen tasks. In this\npaper, we introduce HyGen, a novel hybrid MARL framework, Hybrid Training for\nEnhanced Multi-Task Generalization, which integrates online and offline\nlearning to ensure both multi-task generalization and training efficiency.\nSpecifically, our framework extracts potential general skills from offline\nmulti-task datasets. We then train policies to select the optimal skills under\nthe centralized training and decentralized execution paradigm (CTDE). During\nthis stage, we utilize a replay buffer that integrates both offline data and\nonline interactions. We empirically demonstrate that our framework effectively\nextracts and refines general skills, yielding impressive generalization to\nunseen tasks. Comparative analyses on the StarCraft multi-agent challenge show\nthat HyGen outperforms a wide range of existing solely online and offline\nmethods.","PeriodicalId":501315,"journal":{"name":"arXiv - CS - Multiagent Systems","volume":"43 1","pages":""},"PeriodicalIF":0.0000,"publicationDate":"2024-08-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Hybrid Training for Enhanced Multi-task Generalization in Multi-agent Reinforcement Learning\",\"authors\":\"Mingliang Zhang, Sichang Su, Chengyang He, Guillaume Sartoretti\",\"doi\":\"arxiv-2408.13567\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"In multi-agent reinforcement learning (MARL), achieving multi-task\\ngeneralization to diverse agents and objectives presents significant\\nchallenges. Existing online MARL algorithms primarily focus on single-task\\nperformance, but their lack of multi-task generalization capabilities typically\\nresults in substantial computational waste and limited real-life applicability.\\nMeanwhile, existing offline multi-task MARL approaches are heavily dependent on\\ndata quality, often resulting in poor performance on unseen tasks. In this\\npaper, we introduce HyGen, a novel hybrid MARL framework, Hybrid Training for\\nEnhanced Multi-Task Generalization, which integrates online and offline\\nlearning to ensure both multi-task generalization and training efficiency.\\nSpecifically, our framework extracts potential general skills from offline\\nmulti-task datasets. We then train policies to select the optimal skills under\\nthe centralized training and decentralized execution paradigm (CTDE). During\\nthis stage, we utilize a replay buffer that integrates both offline data and\\nonline interactions. We empirically demonstrate that our framework effectively\\nextracts and refines general skills, yielding impressive generalization to\\nunseen tasks. Comparative analyses on the StarCraft multi-agent challenge show\\nthat HyGen outperforms a wide range of existing solely online and offline\\nmethods.\",\"PeriodicalId\":501315,\"journal\":{\"name\":\"arXiv - CS - Multiagent Systems\",\"volume\":\"43 1\",\"pages\":\"\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2024-08-24\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"arXiv - CS - Multiagent Systems\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/arxiv-2408.13567\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"arXiv - CS - Multiagent Systems","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/arxiv-2408.13567","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
Hybrid Training for Enhanced Multi-task Generalization in Multi-agent Reinforcement Learning
In multi-agent reinforcement learning (MARL), achieving multi-task
generalization to diverse agents and objectives presents significant
challenges. Existing online MARL algorithms primarily focus on single-task
performance, but their lack of multi-task generalization capabilities typically
results in substantial computational waste and limited real-life applicability.
Meanwhile, existing offline multi-task MARL approaches are heavily dependent on
data quality, often resulting in poor performance on unseen tasks. In this
paper, we introduce HyGen, a novel hybrid MARL framework, Hybrid Training for
Enhanced Multi-Task Generalization, which integrates online and offline
learning to ensure both multi-task generalization and training efficiency.
Specifically, our framework extracts potential general skills from offline
multi-task datasets. We then train policies to select the optimal skills under
the centralized training and decentralized execution paradigm (CTDE). During
this stage, we utilize a replay buffer that integrates both offline data and
online interactions. We empirically demonstrate that our framework effectively
extracts and refines general skills, yielding impressive generalization to
unseen tasks. Comparative analyses on the StarCraft multi-agent challenge show
that HyGen outperforms a wide range of existing solely online and offline
methods.