Machine Learning Class (Session #17)

Опубликовано 8 июля 2016, 1:12

October 5: Modeling Day 9:30am-10:30am Model Based Machine Learning 1: A Gentle Introduction Chris Bishop In the traditional approach to problem solving with machine learning, the developer typically selects from amongst the many machine learning algorithms developed over the last few decades, perhaps influenced by their familiarity with specific algorithms and by the availability of corresponding software implementations. In these lectures we describe an alternative methodology for applying machine learning, in which a bespoke solution is formulated for each new application. The solution is expressed through a compact modelling language, and the corresponding custom machine learning code is then generated automatically. This model-based approach offers several advantages, including the opportunity to create highly tailored models for specific scenarios, rapid prototyping and comparison of a range of alternative models, and transparency of functionality. This first lecture on model-based machine learning will introduce the basic concepts of probabilistic models and factor graphs using simple examples, and will show how they can provide the foundation for model-based machine learning. The lecture assumes no previous knowledge of machine learning or probabilities.