Academic
Publications
Planning to Be Surprised: Optimal Bayesian Exploration in Dynamic Environments

Planning to Be Surprised: Optimal Bayesian Exploration in Dynamic Environments,10.1007/978-3-642-22887-2_5,Computing Research Repository,Yi Sun,Fausti

Planning to Be Surprised: Optimal Bayesian Exploration in Dynamic Environments   (Citations: 2)
BibTex | RIS | RefWorks Download
To maximize its success, an AGI typically needs to explore its initially unknown world. Is there an optimal way of doing so? Here we derive an affirmative answer for a broad class of environments.
Journal: Computing Research Repository - CORR , vol. abs/1103.5, 2011
Cumulative Annual
View Publication
The following links allow you to view full publications. These links are maintained by other sources not affiliated with Microsoft Academic Search.
    • ... where the agent chooses at each time step the action maximizing ¯ g (a� h); and iv) a dynamic-programming (DP) approximation of the optimal Bayesian exploration, where at each time step the agent follows a policy which is computed using policy iteration, assuming that the dynamics of the MDP is given by the current posterior, and the reward is the expected information gain ¯ g (a� h). The detail of this algorithm is described in ...

    Yi Sunet al. Planning to Be Surprised: Optimal Bayesian Exploration in Dynamic Envi...

    • ...Another potentially serious over-simplificati on the RL task is reduced to maximizing the expected intrinsic reward at the next time step, rather than using the expected future discounted reward [19]...

    Matthew Luciwet al. Artificial curiosity with planning for autonomous perceptual and cogni...

Sort by: