From Exploration to Planning Cornelius Weber and Jochen Triesch Frankfurt Institute for Advanced Studies Goethe University Frankfurt, Germany 18 th International.

From Exploration to Planning

Cornelius Weber and Jochen TrieschFrankfurt Institute for Advanced StudiesGoethe University Frankfurt, Germany

18th International Conference on Artificial Neural Networks3d - 6th September 2008, Prague

Reinforcement Learning

value actor units

fixed reactive system that always strives for the same goal

Trained Weights

reinforcement learning does not use the exploration phase

to learn a general model of the environment

that would allow the agent to plan a route to any goal

so let’s do this

Learning

actor

state space

randomly move aroundthe state space

learn world models:● associative model● inverse model● forward model

Learning: Associative Model

weights to associateneighbouring states

use these to find any possible routes between agent and goalj

ss'iji sw=' s~ jii

ss'ij s''sε=Δw s~

Learning: Inverse Model

weights to “postdict”action given state pair

use these to identify the action that leads to a desired stateji

s s'akijk s'sw=a ~ jikk

sas'kij s'saaε=Δw ~

sum product Sigma-Pi neuron model

Learning: Forward Model

weights to predict stategiven state-action pair

use these to predict the next state given the chosen actionjk

ass'ikji saw=' s jkii

ass'ikj sa''sε=Δw s

Planning

Planning

Planning

Planning

Planning

Planning

Planning

Planning

Planning

Planning

Planning

Planning

Planning

Planning

Planning

goal

actorunits

agent

Planning

Planning

Planning

Discussion

- reinforcement learning ... if no access to full state space

- previous work ... AI-like planners assume links between states

- noise ... wide “goal hills” will have flat slopes

- shortest path ... not taken; how to define?

- biological plausibility ... Sigma-Pi neurons; winner-take-all

- to do: embedding ... learn state space from sensor input

- to do: embedding ... let the goal be assigned naturally

- to do: embedding ... hand-designed planning phases

Acknowledgments

Collaborators:

Jochen Triesch FIAS J-W-Goethe University Frankfurt

Stefan Wermter University of Sunderland

Mark Elshaw University of Sheffield