Machine Learning Summer School 2005 - Canberra

How to predict with Bayes, MDL, and Experts

author: Marcus Hutter, IDSIA

Description

Most passive Machine Learning tasks can be (re)stated as sequence prediction problems. This includes pattern recognition, classification, time-series forecasting, and others. Moreover, the understanding of passive intelligence also serves as a basis for active learning and decision making. In the recent past, rich theories for sequence prediction have been developed, and this is still an ongoing process. On the other hand, we are arriving at the stage where some important results are already termed classical. While much of the current Learning Theory is formulated under the assumption of independent and identically distributed (i.i.d.) observations, this lecture series focusses on situations without this prerequisite (e.g. weather or stock-market time-series).

Categories

Top: Computer Science: Machine Learning: Statistical Learning
Top: Computer Science: Machine Learning: Computational Learning Theory

You might be experiencing some problems with Your Video player.

*Slides*
0:44	How to Predict with Bayes, MDL, and Experts
1:24	Overview
7:59	Table of Contents
8:44	Philosophical Issues: Contents
9:21	Philosophical Issues: Abstract
9:22	On the Foundations of Machine Learning
11:50	Example 1: Probability of Sunrise Tomorrow
15:29	Example 2: Digits of a Computable Number
17:15	Example 3: Number Sequences
20:57	Occam's Razor to the Rescue
21:58	Foundations of Induction
23:22	Problem Setup
25:15	Dichotomies in Machine Learning
27:53	Sequential/online predictions
30:16	Bayesian Sequence Prediction: Contents
31:11	Bayesian Sequence Prediction: Abstract
31:13	Uncertainty and Probability
32:01	Frequency Interpretation: Counting
33:12	Objective Interpretation: Uncertain Events
35:01	Subjective Interpretation: Degrees of Belief
37:31	Bayes' Famous Rule
41:14	Example: Bayes' and Laplace's Rule

Lecture rating

People found this lecture:
Worth seeing
because it is:
Valuable and informative
Well presented
Easily understandable
Acceptably recorded
You need to login to cast your vote.

Report a problem or upload files

If you have found a problem with this lecture or would like to send us extra material, articles, exercises, etc., please use our ticket system to describe your request and upload the data.
Enter your e-mail into the 'Cc' field, and we will keep you updated with your request's status.