Current browse context:
stat.ML
Change to browse by:
References & Citations
Statistics > Machine Learning
Title: Distributed Multitask Learning
(Submitted on 2 Oct 2015)
Abstract: We consider the problem of distributed multi-task learning, where each machine learns a separate, but related, task. Specifically, each machine learns a linear predictor in high-dimensional space,where all tasks share the same small support. We present a communication-efficient estimator based on the debiased lasso and show that it is comparable with the optimal centralized method.
Link back to: arXiv, form interface, contact.