Filters








8,939 Hits in 3.8 sec

A Generalized Training Approach for Multiagent Learning [article]

Paul Muller, Shayegan Omidshafiei, Mark Rowland, Karl Tuyls, Julien Perolat, Siqi Liu, Daniel Hennes, Luke Marris, Marc Lanctot, Edward Hughes, Zhe Wang, Guy Lever, Nicolas Heess (+2 others)
2020 arXiv   pre-print
This paper investigates a population-based training regime based on game-theoretic principles called Policy-Spaced Response Oracles (PSRO).  ...  We also carry out an initial empirical validation in MuJoCo soccer, illustrating the feasibility of the proposed approach in another complex domain.  ...  ACKNOLWEDGEMENTS The authors gratefully thank Bart De Vylder for providing helpful feedback on the paper draft.  ... 
arXiv:1909.12823v2 fatcat:mt7t6xeiebcgzeqrpug2kszf3y

Scalable multiagent learning through indirect encoding of policy geometry

David B. D'Ambrosio, Kenneth O. Stanley
2013 Evolutionary Intelligence  
While there are a variety of traditional approaches to multiagent learning, many suffer from increased computational costs for large teams and the problem of reinvention (that is, the inability to recognize  ...  In this paper, multiagent HyperNEAT is compared to a traditional learning method, multiagent Sarsa(λ ), in a predator-prey domain, where it demonstrates its ability to train large teams.  ...  That way, this approach to scaling can still be a good starting heuristic for additional training.  ... 
doi:10.1007/s12065-012-0086-3 fatcat:7ropvv3r5nemlh4hkliqeymqgi

Multiagent reactive plan application learning in dynamic environments

Hüseyin Sevay, Costas Tsatsoulis
2002 Proceedings of the first international joint conference on Autonomous agents and multiagent systems part 2 - AAMAS '02  
In addition to bottom-up learning approaches, which facilitate emergent policy learning, it also is desirable to have top-down control over learning so that a team of agents can also learn to apply general  ...  We present a multiagent case-based learning methodology to achieve this top-down control.  ...  In our approach, agents start with a small set of handdesigned multiagent plans without any application knowledge.  ... 
doi:10.1145/544862.544937 dblp:conf/atal/SevayT02 fatcat:vyi5ohkg7nb23and7gwvlxnphq

Multiagent reactive plan application learning in dynamic environments

Hüseyin Sevay, Costas Tsatsoulis
2002 Proceedings of the first international joint conference on Autonomous agents and multiagent systems part 2 - AAMAS '02  
In addition to bottom-up learning approaches, which facilitate emergent policy learning, it also is desirable to have top-down control over learning so that a team of agents can also learn to apply general  ...  We present a multiagent case-based learning methodology to achieve this top-down control.  ...  In our approach, agents start with a small set of handdesigned multiagent plans without any application knowledge.  ... 
doi:10.1145/544932.544937 fatcat:uyc3ztzhzfdy3ixc5hi6do2xom

Generative encoding for multiagent learning

David B. D'Ambrosio, Kenneth O. Stanley
2008 Proceedings of the 10th annual conference on Genetic and evolutionary computation - GECCO '08  
This paper argues that multiagent learning is a potential "killer application" for generative and developmental systems (GDS) because key challenges in learning to coordinate a team of agents are naturally  ...  For example, a significant problem for multiagent learning is that policies learned separately for different agent roles may nevertheless need to share a basic skill set, forcing the learning algorithm  ...  While several approaches to learning have been applied to multiagent systems [11, 27] , this paper argues that multiagent learning may be a "killer application" for GDS.  ... 
doi:10.1145/1389095.1389256 dblp:conf/gecco/DAmbrosioS08 fatcat:nqvwapuiafg7dijcs6kpvoxetu

Beyond Rewards: a Hierarchical Perspective on Offline Multiagent Behavioral Analysis [article]

Shayegan Omidshafiei, Andrei Kapishnikov, Yannick Assogba, Lucas Dixon, Been Kim
2022 arXiv   pre-print
We introduce a model-agnostic method for discovery of behavior clusters in multiagent domains, using variational inference to learn a hierarchy of behaviors at the joint and local agent levels.  ...  In this paper we take a step back from performance-focused multiagent learning, and instead turn our attention towards agent behavior analysis.  ...  Acknowledgments and Disclosure of Funding We thank Asma Ghandeharioun, Meredith Morris, and Kathy Meier-Hellstern for their feedback during the paper writing process.  ... 
arXiv:2206.09046v1 fatcat:utfuxno6vzhvlj2sljiwkaohke

Task switching in multirobot learning through indirect encoding

D. B. D'Ambrosio, J. Lehman, S. Risi, K. O. Stanley
2011 2011 IEEE/RSJ International Conference on Intelligent Robots and Systems  
Multirobot domains are a challenge for learning algorithms because they require robots to learn to cooperate to achieve a common goal.  ...  Multiagent HyperNEAT is a neuroevolutionary method (i.e. a method that evolves neural networks) that has proven successful in several cooperative multiagent domains by exploiting the concept of policy  ...  A. Traditional Cooperative Multiagent Learning There are two primary traditional approaches to multiagent learning.  ... 
doi:10.1109/iros.2011.6094509 dblp:conf/iros/DAmbrosioLRS11 fatcat:6j364cgeyzegpj23vhmclki2xi

A Survey and Critique of Multiagent Deep Reinforcement Learning [article]

Pablo Hernandez-Leal, Bilal Kartal, Matthew E. Taylor
2019 arXiv   pre-print
The primary goal of this article is to provide a clear overview of current multiagent deep reinforcement learning (MDRL) literature.  ...  Recent works have explored learning beyond single-agent scenarios and have considered multiagent learning (MAL) scenarios.  ...  and three anonymous reviewers whose comments and suggestions increased the quality of this work. 56 The title of this work makes a clear reference to previous seminal MAL works [2, 354] .  ... 
arXiv:1810.05587v2 fatcat:h4ei5zx2xfa7xocktlefjrvef4

Task switching in multirobot learning through indirect encoding

David B. D'Ambrosio, Joel Lehman, Sebastian Risi, Kenneth O. Stanley
2011 2011 IEEE/RSJ International Conference on Intelligent Robots and Systems  
A. Traditional Cooperative Multiagent Learning There are two primary traditional approaches to multiagent learning.  ...  Thus, rather than learning individual policies for each agent, the approach of multiagent HyperNEAT is to learn a pattern of policies and how they relate to one another based on the team's policy geometry  ... 
doi:10.1109/iros.2011.6048150 fatcat:36gratlelnfqrmmwdlllrkiray

A multiagent architecture for concurrent reinforcement learning

Victor Uc Cetina
2006 The European Symposium on Artificial Neural Networks  
In this paper we propose a multiagent architecture for implementing concurrent reinforcement learning, an approach where several agents, sharing the same environment, perceptions and actions, work towards  ...  one only objective: learning a single value function.  ...  Architecture Fig. 1: A multiagent architecture for concurrent reinforcement learning.  ... 
dblp:conf/esann/Cetina06 fatcat:vfank2hojvcitcnfawawn2fewi

From Few to More: Large-scale Dynamic Multiagent Curriculum Learning [article]

Weixun Wang, Tianpei Yang, Yong Liu, Jianye Hao, Xiaotian Hao, Yujing Hu, Yingfeng Chen, Changjie Fan, Yang Gao
2019 arXiv   pre-print
In this paper, we design a novel Dynamic Multiagent Curriculum Learning (DyMA-CL) to solve large-scale problems by starting from learning on a multiagent scenario with a small size and progressively increasing  ...  Experimental results show that DyMA-CL using DyAN greatly improves the performance of large-scale multiagent learning compared with state-of-the-art deep reinforcement learning approaches.  ...  In Curriculum Learning, the goal is to generate a series of training tasks, beginning from the simplest one and then gradually increasing the difficulty of training to improve the final asymptotic performance  ... 
arXiv:1909.02790v2 fatcat:fsww4pjwijg6pihi66sphs5fni

Machine Learning Techniques for MultiAgent Systems

Yoad Lewenberg
2017 Proceedings of the Twenty-Sixth International Joint Conference on Artificial Intelligence  
One example of a successful combination is the intersection of machine learning and multiagent systems.  ...  Research in artificial intelligence ranges over many subdisciplines, such as Natural Language Processing, Computer Vision, Machine Learning, and MultiAgent Systems.  ...  The use of multiagent reinforcement learning algorithms has attracted attention because of its generality and robustness.  ... 
doi:10.24963/ijcai.2017/752 dblp:conf/ijcai/Lewenberg17 fatcat:3quytmmecvgl7fpt5yhqfwffzi

From Few to More: Large-Scale Dynamic Multiagent Curriculum Learning

Weixun Wang, Tianpei Yang, Yong Liu, Jianye Hao, Xiaotian Hao, Yujing Hu, Yingfeng Chen, Changjie Fan, Yang Gao
2020 PROCEEDINGS OF THE THIRTIETH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE AND THE TWENTY-EIGHTH INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE CONFERENCE  
In this paper, we design a novel Dynamic Multiagent Curriculum Learning (DyMA-CL) to solve large-scale problems by starting from learning on a multiagent scenario with a small size and progressively increasing  ...  Experimental results show that DyMA-CL using DyAN greatly improves the performance of large-scale multiagent learning compared with state-of-the-art deep reinforcement learning approaches.  ...  In Curriculum Learning, the goal is to generate a series of training tasks, beginning from the simplest one and then gradually increasing the difficulty of training to improve the final asymptotic performance  ... 
doi:10.1609/aaai.v34i05.6221 fatcat:7kwnnbppcrgvbcrmck4mopgviq

Towards collaborative and adversarial learning: a case study in robotic soccer

PETER STONE, MANUELA VELOSO
1998 International Journal of Human-Computer Studies  
Soccer is a rich domain for the study of multiagent learning issues.  ...  We present a learned, robust, low-level behavior that is necessitated by the multiagent nature of the domain, namely shooting a moving ball.  ...  Introduction Soccer is a rich domain for the study of multiagent learning issues.  ... 
doi:10.1006/ijhc.1997.0162 fatcat:vr7iwy4ujjebnhzvjajzuzyazy

Multiage Instruction: An Outdated Strategy, or a Timeless Best Practice

Myron Eighmy
2012 European Journal of Social & Behavioural Sciences  
This investigation provided a foundation of knowledge on multiage instruction regarding strategies and challenges, the pros and cons of multiage instruction, and training and resources needed for the successful  ...  in order for it to be a successful practice.  ...  training and conference experiences for multiage teachers.  ... 
doi:10.15405/futureacademy/ejsbs(2301-2218).2012.2.4 fatcat:4aa5efp53jg5bghapdyqa2s5uq
« Previous Showing results 1 — 15 out of 8,939 results