Variable Selection and Task Grouping for Multi-Task Learning
This work addresses multi-task learning for researchers and practitioners by offering an incremental improvement in model efficiency and interpretability through structured sparsity.
The paper tackles the problem of improving generalization in multi-task learning by proposing a method that factorizes the coefficient matrix with sparsity constraints to perform variable selection and learn overlapping task group structures, validated on synthetic and real-world datasets.
We consider multi-task learning, which simultaneously learns related prediction tasks, to improve generalization performance. We factorize a coefficient matrix as the product of two matrices based on a low-rank assumption. These matrices have sparsities to simultaneously perform variable selection and learn and overlapping group structure among the tasks. The resulting bi-convex objective function is minimized by alternating optimization where sub-problems are solved using alternating direction method of multipliers and accelerated proximal gradient descent. Moreover, we provide the performance bound of the proposed method. The effectiveness of the proposed method is validated for both synthetic and real-world datasets.