Neil D. Lawrence, Department of Computer Science, University of Sheffield
John C. Platt, CCSP Group, Microsoft Research
International Conference on Machine Learning, Paper 65, (2004)
This paper describes an efficient method for learning the parameters of a Gaussian process (GP). The parameters are learning from multiple tasks which are assumed to have been drawn independently from the same GP prior. An efficient algorithm is obtained by extending the informative vector machine (IVM) algorithm to handle the multi-task learning case. The multi-task IVM (MT-IVM) saves computation by greedily selecting the most informative examples from the separate tasks. The MT-IVM is also shown to be more efficient than random sub-sampling on an artificial data-set and more effective than the traditional IVM in a speaker-dependent phoneme recognition task.
PS file (111KB)
The phoneme data set mentioned in the paper is the Deterding Vowel dataset, available at http://www.ics.uci.edu/~mlearn/databases/undocumented/connectionist-bench/vowel/