Can machines learn? Find out at PASCAL Challenges Workshop

A workshop which will investigate the learning power of machines and report advances in this field will be hosted by the University of Southampton from 11 to 13 April 2005. According to Professor John Shawe-Taylor from the University’s School of Electronics and Computer Science (ECS) and organiser of the event, machine learning first became popular in the 1960s, but there has been a resurgence of interest in the last 10 years due to the quality of the new algorithms available. The range of problems that can now be solved using the approach has increased very significantly. Professor Shawe-Taylor showed last year how the complex and individual performance styles of concert pianists such as Glenn Gould and Vladimir Horowitz could be modelled in unique ‘performance alphabets’, providing a method of recognizing their performance styles by computer, and also, possibly, reconstructing them. Now at the Southampton Challenges Workshop, PASCAL – Pattern Analysis, Statistical Modelling and Computation Learning, further examples of machine learning will be demonstrated as the results are presented of four challenge projects set over the last year. The best relevant contributions will be presented from challenges such as asking the machine to recognise and categorise objects; and to decide, given two text fragments whether the meaning of one text can be inferred from another. There will also be feedback on the current situation concerning machine-learning algorithms for information extraction from documents. Professor Shawe-Taylor commented: ‘PASCAL seeks to generate challenges and those who attend the workshop can expect to see state-of-the art machine learning and to get an impression of what these advances can do for them.’