Abstract | ||
---|---|---|
In this paper we are concerned with multi- task learning when task-specific features are available. We describe two ways of achiev- ing this using Gaussian process predictors: in the first method, the data from all tasks is combined into one dataset, making use of the task-specific features. In the second method we train specific predictors for each reference task, and then combine their predictions us- ing a gating network. We demonstrate these methods on a compiler performance predic- tion problem, where a task is defined as pre- dicting the speed-up obtained when applying a sequence of code transformations to a given program. |
Year | Venue | Keywords |
---|---|---|
2007 | AISTATS | gaussian process,multi task learning |
Field | DocType | Volume |
Online machine learning,Semi-supervised learning,Instance-based learning,Multi-task learning,Computer science,Tree kernel,Unsupervised learning,Artificial intelligence,Kernel method,Machine learning,Learning classifier system | Journal | 2 |
Citations | PageRank | References |
35 | 2.87 | 12 |
Authors | ||
3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Edwin V. Bonilla | 1 | 1008 | 53.32 |
Felix V. Agakov | 2 | 442 | 34.22 |
Christopher K. I. Williams | 3 | 6807 | 631.16 |