Convex Multi-Task Learning by Clustering

Abstract

We consider the problem of multi-task learning in which tasks belong to hidden clusters. We formulate the learning problem as a novel convex optimization problem in which linear classifiers are combinations of (a small number of) some basis. Our formulation jointly learns both the basis and the linear combination. We propose a scalable optimization algorithm for finding the optimal solution. Our new methods outperform existing state-of-the-art methods on multi-task sentiment classification tasks.

PDF AISTATS Semantic Scholar

Cite

Text

Barzilai and Crammer. "Convex Multi-Task Learning by Clustering." International Conference on Artificial Intelligence and Statistics, 2015.

Markdown

[Barzilai and Crammer. "Convex Multi-Task Learning by Clustering." International Conference on Artificial Intelligence and Statistics, 2015.](https://mlanthology.org/aistats/2015/barzilai2015aistats-convex/)

BibTeX

@inproceedings{barzilai2015aistats-convex,
  title     = {{Convex Multi-Task Learning by Clustering}},
  author    = {Barzilai, Aviad and Crammer, Koby},
  booktitle = {International Conference on Artificial Intelligence and Statistics},
  year      = {2015},
  url       = {https://mlanthology.org/aistats/2015/barzilai2015aistats-convex/}
}