Bio-Plausible Reinforcement Learning Systems Learn to Play Atari From Human
نویسنده
چکیده
We explore a biologically plausible deep reinforcement learning system by feeding it the human observations of the experiment world. The main hypothesis is that the more similar our learning model with the actual human learning model is, the better the performance should be. We examine this idea by using the AuGMEnT deep neural network which is a bio-plausible reinforcement system with a focus on attention and show that we can instruct our agent the general policies of the environment with just a few episodes of human actions in that world. In addition, we experiment one non-bio-plausible learning system and show that it cannot earn the abilities that our bio-plausible method earns under the same settings.
منابع مشابه
Vision-based Deep Reinforcement Learning
Recently, Google Deepmind showcased how Deep learning can be used in conjunction with existing Reinforcement Learning (RL) techniques to play Atari games[11], beat a world-class player [14] in the game of Go and solve complicated riddles [3]. Deep learning has been shown to be successful in extracting useful, nonlinear features from high-dimensional media such as images, text, video and audio [...
متن کاملPlaying Games with Deep Reinforcement Learning
Recently, Google Deepmind showcased how Deep learning can be used in conjunction with existing Reinforcement Learning (RL) techniques to play Atari games[10], beat a world-class player [13] in the game of Go and solve complicated riddles [3]. Deep learning has been shown to be successful in extracting useful, nonlinear features from high-dimensional media such as images, text, video and audio [...
متن کاملPlaying Atari with Deep Reinforcement Learning
We present the first deep learning model to successfully learn control policies directly from high-dimensional sensory input using reinforcement learning. The model is a convolutional neural network, trained with a variant of Q-learning, whose input is raw pixels and whose output is a value function estimating future rewards. We apply our method to seven Atari 2600 games from the Arcade Learnin...
متن کاملPlaying Atari Games with Deep Reinforcement Learning and Human Checkpoint Replay
This paper introduces a novel method for learning how to play the most difficult Atari 2600 games from the Arcade Learning Environment using deep reinforcement learning. The proposed method, called human checkpoint replay, consists in using checkpoints sampled from human gameplay as starting points for the learning process. This is meant to compensate for the difficulties of current exploration...
متن کاملDeep Reinforcement Learning with POMDPs
Recent work has shown that Deep Q-Networks (DQNs) are capable of learning human-level control policies on a variety of different Atari 2600 games [1]. Other work has looked at treating the Atari problem as a partially observable Markov decision process (POMDP) by adding imperfect state information through image flickering [2]. However, these approaches leverage a convolutional network structure...
متن کامل