Multi-task Supervised Learning via Cross-learning

TL;DR

This paper introduces a multi-task learning framework where task-specific models are coupled to facilitate cross-learning, improving performance across domains through shared information, with algorithms designed for efficiency and empirical validation on image classification tasks.

Contribution

The paper proposes a novel cross-learning approach for multi-task models, coupling parameters to enhance learning across domains, and develops efficient algorithms for large-scale problems.

Findings

Cross-learning outperforms task-specific models.

Algorithms are scalable for large parameter spaces.

Empirical results show improved image classification accuracy.

Abstract

In this paper we consider a problem known as multi-task learning, consisting of fitting a set of classifier or regression functions intended for solving different tasks. In our novel formulation, we couple the parameters of these functions, so that they learn in their task specific domains while staying close to each other. This facilitates cross-fertilization in which data collected across different domains help improving the learning performance at each other task. First, we present a simplified case in which the goal is to estimate the means of two Gaussian variables, for the purpose of gaining some insights on the advantage of the proposed cross-learning strategy. Then we provide a stochastic projected gradient algorithm to perform cross-learning over a generic loss function. If the number of parameters is large, then the projection step becomes computationally expensive. To avoid this situation, we derive a primal-dual algorithm that exploits the structure of the dual problem, achieving a formulation whose complexity only depends on the number of tasks. Preliminary numerical experiments for image classification by neural networks trained on a dataset divided in different domains corroborate that the cross-learned function outperforms both the task-specific and the consensus approaches.