Closing the learning-planning loop with predictive state representations (Extended Abstract)

DMCA

Closing the learning-planning loop with predictive state representations (Extended Abstract) (2010)

Cached

Download Links

by Byron Boots , Sajid M. Siddiqi , Geoffrey J. Gordon

Citations:

49 - 12 self

BibTeX

@MISC{Boots10closingthe,
    author = {Byron Boots and Sajid M. Siddiqi and Geoffrey J. Gordon},
    title = {Closing the learning-planning loop with predictive state representations (Extended Abstract)},
    year = {2010}
}

OpenURL

Abstract

A central problem in artificial intelligence is to plan to maximize future reward under uncertainty in a partially observable environment. Models of such environments include Partially Observable Markov Decision Processes (POMDPs) [4] as well as their generalizations, Predictive State Representations (PSRs) [9] and Observable Operator Models (OOMs) [7]. POMDPs model the state of the world as a latent variable; in contrast, PSRs and OOMs represent state by tracking occurrence probabilities of a set of future events (called tests or characteristic events) conditioned on past events (called histories or indicative events). Unfortunately, exact planning algorithms such as value iteration [14] are intractable for most realistic POMDPs due to the curse of history and the curse of dimensionality [11]. However, PSRs and OOMs hold the promise of mitigating both of these curses: first, many successful approximate planning techniques designed to address

Keyphrases

general term central problem characteristic event predictive state representation future event future reward partially observable markov decision process many successful approximate planning technique indicative event latent variable occurrence probability observable environment observable operator model past event value iteration realistic pomdps artificial intelligence

DMCA