This course presents advanced materials on datamining, pattern recognition, and machine learning

A brief outline of the course:

Machine learning techniques are widely used in many computing applications; for example, in web search engines, spam filtering, speech and image recognition, computer games, machine vision, credit card fraud detection, stock market analysis and product marketing applications. Machine learning implies that there is some improvement that results from the learning program having seen some data. The improvement can be in terms of some performance program (e.g., learning an expert system or improving the performance of a planning or scheduling program), in terms of finding an unknown relation in the data (e.g., data mining, pattern analysis), or in terms of customizing adaptive systems (e.g., adaptive user-interfaces or adaptive agents).

In the first half of the paper, we will provide an overview of the learning problem and the view of learning as search. We will study several techniques for learning such as Rule Learning, Exhaustive Learning, Genetic Algorithms, Reinforcement Learning, Neural Networks, and Inductive Logic Programming. In addition we will provide an overview of the experimental methods necessary for understanding machine learning research.

The second half of the paper be deal with advanced machine learning and discovery. This will include applications to recommender systems and other topics, learning in the context of problem solving and planning, and computational discovery of scientific knowledge. The latter two topics involve operating over richer representations than mainstream machine learning.

For lecture slides / handouts - go to Lectures tab.

Required text: