Modular Universal Reparameterization: Deep Multi-task Learning Across Diverse Domains
arXiv:1906.00097
Abstract
As deep learning applications continue to become more diverse, an interesting question arises: Can general problem solving arise from jointly learning several such diverse tasks? To approach this question, deep multi-task learning is extended in this paper to the setting where there is no obvious overlap between task architectures. The idea is that any set of (architecture,task) pairs can be decomposed into a set of potentially related subproblems, whose sharing is optimized by an efficient stochastic algorithm. The approach is first validated in a classic synthetic multi-task learning benchmark, and then applied to sharing across disparate architectures for vision, NLP, and genomics tasks. It discovers regularities across these domains, encodes them into sharable modules, and combines these modules systematically to improve performance in the individual tasks. The results confirm that sharing learned functionality across diverse domains and architectures is indeed beneficial, thus establishing a key ingredient for general problem solving in the future.
arXiv comment33rd Conference on Neural Information Processing Systems (NeurIPS 2019), 16 pages, including Supplemental Material
References in corpus (11)
- Recurrent Neural Network Regularization
- Wide Residual Networks
- Theoretical Models of Learning to Learn
- Learning multiple visual domains with residual adapters
- Reinforcement Learning with Unsupervised Auxiliary Tasks
- Outrageously Large Neural Networks: The Sparsely-Gated Mixture-of-Experts Layer
- One Model To Learn Them All
- Learning Task Grouping and Overlap in Multi-task Learning
- Learning Multiple Tasks with Multilinear Relationship Networks
- Deep Multi-task Representation Learning: A Tensor Factorisation Approach
- Pseudo-task Augmentation: From Deep Multitask Learning to Intratask Sharing---and Back