Mathematical and Machine Learning approaches to Context

3/11/99


Click here to start


Table of Contents

Mathematical and Machine Learning approaches to Context

References for Machine Learning

Machine Learning is often described as Classification

Leaning problems for context-aware applications

Traditional mathematical inference techniques

Watch out for hidden assumptions in machine learning

Generalization Lattice

Version Spaces

Version Spaces

Case-Based Reasoning

Explanation-Based Generalization

Learning by Analogy

Sequence prediction

Approaches that use large numbers of examples

Large-example-set approaches appropriate for context problems?

Statistical approaches: Clustering

Discrimination Net

Feature selection problem

The Frame Problem in AI

Circumscription

Inference rules take "extra argument" of context

Author: Henry Lieberman

Email: lieber@media.mit.edu

Home Page: http://www.media.mit.edu/~lieber

Download presentation source