Distributed Multi-Task Learning with Shared Representation
This addresses the problem of efficient distributed learning for related tasks in multi-machine settings, but appears incremental as it builds on existing multi-task learning concepts.
The paper tackles distributed multi-task learning where each machine learns a separate but related task in a shared low-dimensional subspace, focusing on communication-efficient methods to exploit this shared structure. The result is not specified with concrete numbers in the abstract.
We study the problem of distributed multi-task learning with shared representation, where each machine aims to learn a separate, but related, task in an unknown shared low-dimensional subspaces, i.e. when the predictor matrix has low rank. We consider a setting where each task is handled by a different machine, with samples for the task available locally on the machine, and study communication-efficient methods for exploiting the shared structure.