Filters








58 Hits in 6.2 sec

PIC: Permutation Invariant Critic for Multi-Agent Deep Reinforcement Learning [article]

Iou-Jen Liu, Raymond A. Yeh, Alexander G. Schwing
2019 arXiv   pre-print
Sample efficiency and scalability to a large number of agents are two important goals for multi-agent reinforcement learning systems.  ...  To avoid this inefficiency, we propose a 'permutation invariant critic' (PIC), which yields identical output irrespective of the agent permutation.  ...  We thank NVIDIA for providing GPUs used for this work and Cisco for access to the Arcetri cluster.  ... 
arXiv:1911.00025v1 fatcat:36asva2xfrg4bd6ad4stjgp37a

Permutation Invariant Policy Optimization for Mean-Field Multi-Agent Reinforcement Learning: A Principled Approach [article]

Yan Li, Lingxiao Wang, Jiachen Yang, Ethan Wang, Zhaoran Wang, Tuo Zhao, Hongyuan Zha
2021 arXiv   pre-print
Multi-agent reinforcement learning (MARL) becomes more challenging in the presence of more agents, as the capacity of the joint state and action spaces grows exponentially in the number of agents.  ...  To exploit the permutation invariance therein, we propose the mean-field proximal policy optimization (MF-PPO) algorithm, at the core of which is a permutation-invariant actor-critic neural architecture  ...  Pic: Permutation invariant critic for multi-agent deep reinforcement learning. arXiv preprint arXiv:1911.00025 . Lowe, R., Wu, Y., Tamar, A., Harb, J., Abbeel, O. P. and Mordatch, I. (2017).  ... 
arXiv:2105.08268v1 fatcat:t2ioha7y45covcw6o7uxvmbwdm

API: Boosting Multi-Agent Reinforcement Learning via Agent-Permutation-Invariant Networks [article]

Xiaotian Hao, Weixun Wang, Hangyu Mao, Yaodong Yang, Dong Li, Yan Zheng, Zhen Wang, Jianye Hao
2022 arXiv   pre-print
Multi-agent reinforcement learning suffers from poor sample efficiency due to the exponential growth of the state-action space.  ...  differently ordered representations, thus designing functions satisfying permutation invariant (PI) can reduce the state space by a factor of 1/m!.  ...  Introduction Multi-agent reinforcement learning (MARL) has successfully addressed many complex real-world problems, such as multi-player games (Vinyals et al., 2019; Berner et al., 2019) , autonomous  ... 
arXiv:2203.05285v1 fatcat:prbuiocbrrcnnicqff6ltfsapa

Agent-Temporal Attention for Reward Redistribution in Episodic Multi-Agent Reinforcement Learning [article]

Baicen Xiao, Bhaskar Ramasubramanian, Radha Poovendran
2022 arXiv   pre-print
In this paper, we introduce Agent-Temporal Attention for Reward Redistribution in Episodic Multi-Agent Reinforcement Learning (AREL) to address these two challenges.  ...  This paper considers multi-agent reinforcement learning (MARL) tasks where agents receive a shared global reward at the end of an episode.  ...  In Particle World, we use the permutation invariant critic (PIC) based on MADDPG from [23] as the base reinforcement learning algorithm.  ... 
arXiv:2201.04612v1 fatcat:wwfce65wn5d6niihkqurlvudqu

The Sensory Neuron as a Transformer: Permutation-Invariant Neural Networks for Reinforcement Learning [article]

Yujin Tang, David Ha
2021 arXiv   pre-print
These permutation invariant systems also display useful robustness and generalization properties that are broadly applicable.  ...  In complex systems, we often observe complex global behavior emerge from a collection of agents interacting with each other in their environment, with each individual agent acting only on locally available  ...  Pic: permutation invariant critic for multi-agent deep reinforcement learning. In Conference on Robot Learning, pages 590–602. PMLR, 2020. [50] M.-T. Luong, H. Pham, and C. D. Manning.  ... 
arXiv:2109.02869v2 fatcat:uf4nviubnffz3f2qqui24zc3ei

Variational Policy Propagation for Multi-agent Reinforcement Learning [article]

Chao Qu, Hui Li, Chang Liu, Junwu Xiong, James Zhang, Wei Chu, Weiqiang Wang, Yuan Qi, Le Song
2022 arXiv   pre-print
We propose a collaborative multi-agent reinforcement learning algorithm named variational policy propagation (VPP) to learn a joint policy through the interactions over agents.  ...  ., 2017] ; permutation invariant critic (PIC) ; graph convolutional reinforcement learning (DGN) [Jiang et al., 2020] ; Neurcomm [Chu et al., 2020] ; COMA ; MFQ ; Independent Q learning (IQL) [Tan,  ...  Introduction Collaborative multi-agent reinforcement learning is an important sub-field of the multiagent reinforcement learning (MARL) , where the agents learn to coordinate to achieve joint success.  ... 
arXiv:2004.08883v4 fatcat:d35vb6d4bndmthoq3kxxj3n6la

A Cordial Sync: Going Beyond Marginal Policies for Multi-Agent Embodied Tasks [article]

Unnat Jain, Luca Weihs, Eric Kolve, Ali Farhadi, Svetlana Lazebnik, Aniruddha Kembhavi, Alexander Schwing
2020 arXiv   pre-print
Autonomous agents must learn to collaborate. It is not scalable to develop a new centralized agent every time a task's difficulty outpaces a single agent's abilities.  ...  While multi-agent collaboration research has flourished in gridworld-like environments, relatively little work has considered visually rich domains.  ...  UJ is thankful to Thomas & Stacey Siebel Foundation for Siebel Scholars Award. We thank Mitchell Wortsman and Kuo-Hao Zeng for their insightful suggestions on how to clarify and structure this work.  ... 
arXiv:2007.04979v1 fatcat:2mbgl55d3feizntuqxg37ktovi

Handbook on peace education

Dieter Luense, Mizzi Walker, Renate Grasse, Tina Ottman, Ruth Hayhoe
2013 Journal of Peace Education  
We have also variables related to the groups involved in the power struggle at one level or the other, such as the number of alliances, the number of multi-national corporations, IGOs and NGOs, as well  ...  We have, for example, perceived value capabilities, relative deprivation, frustration, anomie, aggressiveness, perceived hostility, collective violence, image distortion, etc.  ...  It is much more economical than reinforcement learning, which'involves slmpler units of behavior, learned sequentially.  ... 
doi:10.1080/17400201.2013.780862 fatcat:2zupnfkycraglobxukgpi67mmi

Machine learning and the physical sciences [article]

Giuseppe Carleo, Ignacio Cirac, Kyle Cranmer, Laurent Daudet, Maria Schuld, Naftali Tishby, Leslie Vogt-Maranto, Lenka Zdeborová
2019 arXiv   pre-print
Machine learning encompasses a broad range of algorithms and modeling tools used for a vast array of data processing tasks, which has entered most scientific disciplines in recent years.  ...  We review in a selective way the recent research on the interface between machine learning and physical sciences.This includes conceptual developments in machine learning (ML) motivated by physical insights  ...  For bosons, this amounts to imposing the wave-function to be permutationally invariant with respect to exchange of particle indices.  ... 
arXiv:1903.10563v1 fatcat:dk4dqdxfarchnp7i7qj2n6g7fi

Reinforcement Learning (Dagstuhl Seminar 13321) The Critical Internet Infrastructure (Dagstuhl Seminar 13322) Coding Theory (Dagstuhl Seminar 13351) Interaction with Information for Visual Reasoning (Dagstuhl Seminar 13352)

Peter Auer, Marcus Hutter, Laurent, Georg Carle, Jochen Schiller, Steve Uhlig, Walter Willinger, Matthias Wählisch, Thore Husfeldt, Ramamohan Paturi, Gregory Sorkin, Ryan Williams (+14 others)
unpublished
We gratefully acknowledge the Dagstuhl staff for helping on all administrative coordination, for their patience, and most importantly for providing an extremely inspiring environment.  ...  Acknowledgments The editors of this report would like to thank all participants for very fruitful and open-minded discussions! In particular, we thank the operators for sharing practical insights.  ...  In this talk we present PQ-learning, a new Reinforcement Learning (RL)algorithm that determines the rational behaviours of an agent in multi-objective domains.  ... 
fatcat:iz3co5xisfejfbnx2t6ptrorfm

Dagstuhl Reports, Volume 9, Issue 12, December 2019, Complete Issue

2020
AI for Accessibility in Games Tommy Thompson  ...  Deep reinforcement learning is a potential approach for the GVGAI learning track, but has not been investigated yet. Other methods might have potential too.  ...  As for deep learning, now that the early enthusiasm has waned a little, the first criticisms of it, which explain its many limitations, are already being published.  ... 
doi:10.4230/dagrep.9.12 fatcat:hebigxkvinhjdb6qlg3j5hw25u

Abstracts of Working Papers in Economics

1992 Abstracts of Working Papers in Economics  
In this paper we model the determinants of firm level wages and employment explicitly allowing for firm and worker heterogeneity.  ...  We apply the model to a representative sample of 1,097 French enterprises for the period 1978 to 1987. We find that firms with enterprise level agreements appear to implement incentive contracts.  ...  By comparison, inefficiency is 0(l/m) for a dual price mechanism and 0(l/m*.5)) for a fixed price mechanism.  ... 
doi:10.1017/s0951007900002898 fatcat:g2puq3vlwnbdbblxfnancxbwbm

Abstracts Presented at the 2004 Annual Meeting of the Australian Society for the Brain Impairment (ASSBI) and the International Neuropsychological Society (INS), July 7–10, 2004, Brisbane, Australia

2004 Journal of the International Neuropsychological Society  
A model for emotion processing will be presented that considers different routes to perceptual categorization of input and the role of emotional responsivity.  ...  Finally, we will make a case for multi-level assessment for DCD to ensure that children who do have difficulties with motor learning are identified and supported.  ...  The strategy adopted an eclectic approach that integrated errorless learn- ing and multi-sensory environment enrichment concepts.  ... 
doi:10.1017/s1355617704040032 fatcat:vhwnpun3ijaythkza2x4s7es7i

Chapter Three Equalities [chapter]

2016 Everyday Women's and Gender Studies  
Organized around key concepts-from knowledges and identities to representation and places-it provides a flexible format to engage students in the challenges and pleasures of thinking critically about gender  ...  Precisely because it is not the same for everyone, the everyday becomes the ideal location for cultivating students' intellectual capacities as well as their political investigations and interventions.  ...  Perhaps the spark of sexual desire set off deep within one's core reinforces the belief that sexuality is personal, private.  ... 
doi:10.4324/9781315643205-10 fatcat:ikiniciqyfcc5oaslgu2ouwsru

JSE 33:1 Spring 2019 Whole Issue PDF

Kathleen E. Erickson
2019 Journal of Scientific Exploration  
However, the statistical treatment has been criticized and data of the most recent experiments were re-analyzed using proper permutation methods (Grote 2017).  ...  He speculated on the existence of stages of condensation of this agent (for other infra-red absorption tests see Hope et al. 1933) .  ...  AIMS AND SCOPE: The Journal of Scientifi c Exploration publishes material consistent with the Society's mission: to provide a professional forum for critical discussion of topics that are for various reasons  ... 
doaj:29c0d635f69745848ceb506a17ab6322 fatcat:w5unqomi35ci3it4i4bijcdgwa
« Previous Showing results 1 — 15 out of 58 results