reinforcement-learning   1500

« earlier    

Lessons From Alpha Zero (part 6) — Hyperparameter Tuning
This is the sixth installment in our series on lessons learned from implementing AlphaZero. Check out Part 1, Part 2, Part 3, Part 4, and Part 5. In this post, we summarize the configuration and…
machine-learning  reinforcement-learning  games 
6 days ago by jseppanen
[1810.09202] Graph Convolutional Reinforcement Learning for Multi-Agent Cooperation
Learning to cooperate is crucially important in multi-agent environments. The key is to understand the mutual interplay between agents. However, multi-agent environments are highly dynamic, which makes it hard to learn abstract representations of their mutual interplay. In this paper, we propose graph convolutional reinforcement learning for multi-agent cooperation, where graph convolution adapts to the dynamics of the underlying graph of the multi-agent environment, and relation kernels capture the interplay between agents by their relation representations. Latent features produced by convolutional layers from gradually increased receptive fields are exploited to learn cooperation, and the cooperation is further boosted by temporal relation regularization for consistency. Empirically, we show that our method substantially outperforms existing methods in a variety of cooperative scenarios.
machine-learning  agent-based  coordination  rather-interesting  reinforcement-learning  teams  to-understand  rather-convoluted-(heh)  collective-behavior 
6 days ago by Vaguery
[1901.08162] Causal Reasoning from Meta-reinforcement Learning
Discovering and exploiting the causal structure in the environment is a crucial challenge for intelligent agents. Here we explore whether causal reasoning can emerge via meta-reinforcement learning. We train a recurrent network with model-free reinforcement learning to solve a range of problems that each contain causal structure. We find that the trained agent can perform causal reasoning in novel situations in order to obtain rewards. The agent can select informative interventions, draw causal inferences from observational data, and make counterfactual predictions. Although established formal causal reasoning algorithms also exist, in this paper we show that such reasoning can arise from model-free reinforcement learning, and suggest that causal reasoning in complex settings may benefit from the more end-to-end learning-based approaches presented here. This work also offers new strategies for structured exploration in reinforcement learning, by providing agents with the ability to perform -- and interpret -- experiments.
causality  meta-learning  reinforcement-learning 
25 days ago by arsyed
[1901.08162] Causal Reasoning from Meta-reinforcement Learning
Discovering and exploiting the causal structure in the environment is a crucial challenge for intelligent agents. Here we explore whether causal reasoning can emerge via meta-reinforcement learning. We train a recurrent network with model-free reinforcement learning to solve a range of problems that each contain causal structure. We find that the trained agent can perform causal reasoning in novel situations in order to obtain rewards. The agent can select informative interventions, draw causal inferences from observational data, and make counterfactual predictions. Although established formal causal reasoning algorithms also exist, in this paper we show that such reasoning can arise from model-free reinforcement learning, and suggest that causal reasoning in complex settings may benefit from the more end-to-end learning-based approaches presented here. This work also offers new strategies for structured exploration in reinforcement learning, by providing agents with the ability to perform -- and interpret -- experiments.
causality  modeling  machine-learning  reinforcement-learning  side-effects  to-understand 
25 days ago by Vaguery
AlphaStar: Mastering the Real-Time Strategy Game StarCraft II | DeepMind
StarCraft, considered to be one of the most challenging Real-Time Strategy games and one of the longest-played esports of all time, has emerged by consensus as a “grand challenge” for AI research. Here, we introduce our StarCraft II program AlphaStar, the first Artificial Intelligence to defeat a top professional player.
starcraft2  deepmind  reinforcement-learning  deep-learnning 
25 days ago by pmigdal

« earlier    

related tags

agent-based  ai  algorithm  algorithms  also-note-presentation  apm  artificial-intelligence  arxiv  autonomous-vehicles  book  books  causal-inference  causality  collective-behavior  combinatorial-optimization  computer-vision  control-theory  coordination  course  curiosity  deep-learning  deep-learnning  deeplearning  deepmind  dl  doesnt-work  driving  ebooks  ethics  exploration-exploitation  exploration  facebook  finance  framework  game-mod  game-theory  games  google  graphics  hn  horizon  imitation-learning  interactive-fiction  karpathy  learning  machine-learning  machine_learning  machinelearning  meta-learning  ml  modeling  nlp  open-ai  open-source  openai  paper  pdf  pong  prejudice  probability  pytorch  rather-convoluted-(heh)  rather-interesting  reinforcement  reinforcementlearning  risk  rl  robotics  robots  search  side-effects  simulation  software  star-craft  starcraft  starcraft2  study-group  stupidity  teams  tencent  tensorflow  textbook  the-mangle-in-practice  to-understand  to-write-about  toread  trading  tutorial  uber  uncertainty  waymo 

Copy this bookmark:



description:


tags: