Aprendizado por Reforço

Aprendizado por Reforço (1999)

Cached

Download Links

[www.ele.ita.br]

by Carlos Henrique Costa Ribeiro

Venue:

JOINT CONFERENCE ON NEURAL NETWORKS (IJCNN’99)- TUTORIAL TRACK ON LEARNING, COM O TÍTULO A TUTORIAL ON REINFORCEMENT LEARNING TECHNIQUES.

BibTeX

@MISC{Ribeiro99aprendizadopor,
    author = {Carlos Henrique Costa Ribeiro},
    title = { Aprendizado por Reforço },
    year = {1999}
}

Bookmark

OpenURL

Abstract

Aprendizado por Reforço (AR) é aprendizado via experimentação direta. Não assume aexistência de um professor provedor de exemplos a partir do qual o aprendizado se desenvolve. Em AR, a experiência éaúnica professora. Com raizes históricas no estudo de reflexos condicionados, AR logo atraiu o interesse de Engenheiros e Cientistas da Computação por sua relevância teórica e aplicações potenciais em campos tão diversos quanto Pesquisa Operacional e Robótica. Computacionalmente, AR opera em um ambiente de aprendizagem composto por dois elementos: o aprendiz e um processo dinâmico. A passos de tempo sucessivos, o aprendiz faz uma observação do estado de processo, seleciona uma ação e a aplica de volta ao processo. Sua meta é descobrir uma política de ações que controle o comportamento deste processo dinâmico, usando para isso sinais (reforços) indicativos de quão bem está sendo executada a tarefa em questão. Estes sinais normalmente são associados a alguma condição dramática — por exemplo, realização de uma subtarefa (recompensa) ou completo fracasso (castigo), e a meta do aprendiz é aperfeiçoar seu comportamento baseado em uma medida de desempenho (função dos reforços recebidos).

Citations

3116	A tutorial on hidden Markov models and selected applications in speech recognition - Rabiner - 1989
2959	Artificial Intelligence: a Modern Approach - Russell, Norvig - 1995
2715	Neural networks: A comprehensive foundation (2nd ed.). Upper Saddle River - Haykin - 1999
1964	Dynamic Programming - Bellman - 1957
1313	Finding structure in time - Elman - 1990
1135	Learning from delayed rewards - Watkins - 1989
1060	Learning to predict by the methods of temporal differences - Sutton - 1988
545	Some Studies in Machine Learning using the Game of Checkers - Samuel - 2000
447	Adaptive switching circuits - Widrow, Ho€ - 1960
427	Integrated architectures for learning, planning, and reacting based on approximating integrated architectures for learning, planning, and reacting based on approximating dynamic programming - Sutton - 1990
420	A stochastic approximation method - Robbins, Monro - 1951
334	Practical issues in temporal difference learning - Tesauro - 1992
330	Temporal Difference Learning and TDGammon - Tesauro - 1995
316	Automatic programming of behavior-based robots using reinforcement learning - Mahadevan, Connell - 1992
296	Elephants don’t play chess - Brooks - 1990
275	Prioritized sweeping: Reinforcement learning with less data and less time - Moore, Atkeson - 1993
255	Self-improving reactive agents based on reinforcement learning, planning, and teaching - Lin - 1992
243	Reinforcement Learning with Selective Perception and Hidden State - McCallum - 1996
223	Generalization in reinforcement learning: safely approximating the value function - Boyan, Moore - 1995
186	On the convergence of stochastic iterative dynamic programming algorithms - Jaakkola, Jordan, et al. - 1994
185	Learning and sequential decision making - Barto, Sutton - 1990
185	Cognitive systems based on adaptive algorithms - Holland, Reitman - 1977
173	Reinforcement learning with perceptual aliasing: The perceptual distinctions approach - Chrisman - 1992
166	Reinforcement learning with replacing eligibility traces - Singh, Sutton - 1996
157	Neuronlike elements that can solve difficult learning control problems - Barto, Sutton, et al. - 1983
149	Active Vision - Blake, Yuille - 1992
130	A theory of cerebellar function - Albus - 1971
125	A distributed model for mobile robot environment-learning and navigation - Mataric - 1990
119	Feature-based methods for large scale dynamic programming - Tsitsiklis, Roy - 1996
117	input generalization in delayed reinforcement learning: an algorithm and performance comparisons - Chapman, Kaelbling
114	aJld �l. Colombetti. Robot shaping: Developing autonomous agents through learning - Dorigo - 1994
111	BOXES: An experiment in adaptive control - Michie, Chambers - 1968
111	Reinforcement learning for dynamic channel allocation in cellular telephone systems - Singh, Bertsekas - 1997
97	Reinforcement learning with soft state aggregation - Singh, Jaakkola, et al. - 1995
95	Time-derivative models of pavlovian reinforcement - Sutton, Barto - 1990
81	Incremental multi-step Q-learning - Peng, Williams - 1994
80	Dynamic Programming and Optimal Control, volume 1 and 2. Athena Scientific - Bertsekas - 1995
71	Active perception and reinforcement learning - Whitehead, Ballard
59	Memory approaches to reinforcement learning in non-Markovian domains - Lin, Mitchell - 1999
56	Behavior control for robotic exploration of planetary surfaces - Gat, Desai, et al. - 1994
52	CMAC: An associative neural network alternative to backpropagation - Miller, Glanz, et al. - 1990
50	Pattern recognizing control systems - Widrow, Smith - 1964
45	A qualitative approach to robot exploration and map learning - Kuipers - 1987
43	Learning to Solve Markovian Decision Processes - Singh - 1994
32	Self-improving reactive agents:: Case studies of reinforcement learning frameworks - Lin - 1990
32	Markovian Decision Problems - Puterman - 1994
31	Hidden state and reinforcement learning with instancebased state identification - McCallum - 2002
31	Punish/Reward: Learning with a critic in adaptive threshold systems - Widrow, Gupta, et al. - 1973
29	Reinforcement learning of non-Markov decision processes - Whitehead, Lin - 1995
27	M.J.: Real robots, real learning problems - Brooks, Mataric - 1994