Skip to content | Change text size
 

School of Business Systems seminar

26 October 2005

'A Glimpse of PRISM Modelling' by Professor Taisuke Sato, Tokyo Institute of Technology, Japan

When: Tuesday 25 October, 2 - 3 pm
Where: Room 115, building 63, level 1, Clayton campus

PRISM (Programming in Statistical Modelling) is a new type of logic programming language for statistical modelling that can learn from data. It was born out of desire to integrate logical reasoning and probabilistic reasoning and has been continuously developed as a tool for modelling complex phenomena governed by rules and probabilities that go beyond existing frameworks such as Bayesian networks and Probabilistic Context-Free Grammars.

In this talk, Professor Sato will first review PRISM's basic structure, i.e. semantics, tabled search, dynamic programming, and Expectation-Maximization learning of statistical parameters in programs. Then Professor Sato will present the recent development of statistical constraint modelling that combines hidden Markov models and constraints. Professor Sato will show how parameter learning for distributions defined by such models becomes possible by program transformation.

See the PRISM web page.

Further information is available at the Business Systems website or by contacting Ms Daisy Seng on +61 3 9905 9082 or email daisy.seng@infotech.monash.edu.au.