In many real-world problems, reward signals received by agents are delayed or sparse, which makes it challenging to train a reinforcement learning (RL) agent. An intrinsic signal can help an agent explore such environments in the quest for novel states. this work, we propose general end-to-end diversity-augmented motivation deep encourages new states and automatically provides denser rewards. S...

This paper presents an actor-critic deep reinforcement learning agent with experience replay that is stable, sample efficient, and performs remarkably well on challenging environments, including the discrete 57-game Atari domain and several continuous control problems. To achieve this, the paper introduces several innovations, including truncated importance sampling with bias correction, stocha...

Machine learning algorithms for controlling devices will need to learn quickly, with few trials. Such a goal can be attained with concepts borrowed from continental philosophy and formalized using tools from the mathematical theory of categories. Illustrations of this approach are presented on a cyberphysical system: the slot car game, and also on Atari 2600 games.

In 1983, the videogame market in the USA collapsed, leading to bankruptcy for more than 90 percent of game developers, as well as Atari, manufacturer of the dominant game console at the time. The main reason was a ‘lemons’ market failure: because it had not developed a technology for locking out unauthorized games, Atari was unable to prevent the entry of opportunistic developers, who flooded t...

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...

In this issue of Neuron, Cross et al., 2021Cross L. Cockburn J. Yue Y. O’Doherty J.P. Using deep reinforcement learning to reveal how the brain encodes abstract state-space representations in high-dimensional environments.Neuron. 2021; 109 (this issue): 724-738Abstract Full Text PDF Scopus (7) Google Scholar use a algorithm understand human neural activation evoked by playing different video ga...

Abstract Deep reinforcement learning (DRL) requires large samples and a long training time to operate optimally. Yet humans rarely require periods of perform well on novel tasks, such as computer games, once they are provided with an accurate program instructions. We used perceptual control theory (PCT) construct simple closed-loop model which no within video game study using the Arcade Learnin...

The basic CoSMoS process concerns the design, implementation, and use of a simulation built from scratch. However, the CoSMoS approach may be tailored and adapted for other styles of use. Here we describe how it has been applied to analyse and re-engineer an existing simulation, that of Schelling’s Bounded Neighbourhood Model. We find that using a principled approach to the analysis of an exist...