# Lect15 - Reasoning under Uncertainty Extrapolate Shift from...

This preview shows pages 1–5. Sign up to view the full content.

Reasoning under Uncertainty Extrapolate Shift from Classical Planning to Reinforcement Learning Classical Planning Logic as reasoning engine Model is purely analytic Deep (first-order expressiveness) Brittle (uncertainty is unavoidable and catastrophic) Reinforcement Learning Decision Theory (statistics) as reasoning engine Model is largely empirical Shallow (propositional expressiveness) Robust (model is fit / adapted to the real world) Apply this paradigm shift to reasoning more generally

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
Reasoning under Uncertainty Qualification Problem Laziness Theoretical Ignorance Practical Ignorance Computational Issues It is sometimes better not to model the world accurately (consider a coffee cup…) Our models are always “approximate” “All models are wrong, but some are useful” - George Box, famous statistician
Ontology / Semantics of Uncertainty Fuzzy logic the world is imprecise “John is tall” Linguistic variables Characteristic functions / fuzzy sets Frequentist statistics the distribution is real (world) Classical statistics; probability = long run average Data helps us approximate it Bayesian statistics the data is real Subjective probabilities Evidence for different distributions

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
Religious wars among Statisticians (are you a Bayesian?) Frequentist / objectivist / classical statistics / Fisherian
This is the end of the preview. Sign up to access the rest of the document.

## This note was uploaded on 10/13/2011 for the course CS 440 taught by Professor Levinson,s during the Fall '08 term at University of Illinois, Urbana Champaign.

### Page1 / 17

Lect15 - Reasoning under Uncertainty Extrapolate Shift from...

This preview shows document pages 1 - 5. Sign up to view the full document.

View Full Document
Ask a homework question - tutors are online