Pareto Multi-Task Learning

Abstract

Multi-task learning is a powerful method for solving multiple correlated tasks simultaneously. However, it is often impossible to find one single solution to optimize all the tasks, since different tasks might conflict with each other. Recently, a novel method is proposed to find one single Pareto optimal solution with good trade-off among different tasks by casting multi-task learning as multiobjective optimization. In this paper, we generalize this idea and propose a novel Pareto multi-task learning algorithm (Pareto MTL) to find a set of well-distributed Pareto solutions which can represent different trade-offs among different tasks. The proposed algorithm first formulates a multi-task learning problem as a multiobjective optimization problem, and then decomposes the multiobjective optimization problem into a set of constrained subproblems with different trade-off preferences. By solving these subproblems in parallel, Pareto MTL can find a set of well-representative Pareto optimal solutions with different trade-off among all tasks. Practitioners can easily select their preferred solution from these Pareto solutions, or use different trade-off solutions for different situations. Experimental results confirm that the proposed algorithm can generate well-representative solutions and outperform some state-of-the-art algorithms on many multi-task learning applications.

33rd Conference on Neural Information Processing Systems (NeurIPS 2019), Vancouver, Canada

Cited by in corpus (27)

• Multi-Task Learning for Dense Prediction Tasks: A Survey
• Multi-Task Learning with Deep Neural Networks: A Survey
• Distribution Matching for Heterogeneous Multi-Task Learning: a Large-scale Face Study
• Gradient Vaccine: Investigating and Improving Multi-task Optimization in Massively Multilingual Models
• Learning the Pareto Front with Hypernetworks
• Federated Learning Meets Multi-objective Optimization
• Efficiently Identifying Task Groupings for Multi-Task Learning
• Multilingual Speech Translation with Efficient Finetuning of Pretrained Models
• Controllable Pareto Multi-Task Learning
• Random Hypervolume Scalarizations for Provable Multi-Objective Black Box Optimization
• A Simple General Approach to Balance Task Difficulty in Multi-Task Learning
• Measuring and Harnessing Transference in Multi-Task Learning
• Small Towers Make Big Differences
• The Fairness-Accuracy Pareto Front
• Pareto Efficient Fairness in Supervised Learning: From Extraction to Tracing
• Multi-Task Variational Information Bottleneck
• Multi-Domain Multi-Task Rehearsal for Lifelong Learning
• Multi-Objective Learning to Predict Pareto Fronts Using Hypervolume Maximization
• Learning Boost by Exploiting the Auxiliary Task in Multi-task Domain
• Self-Evolutionary Optimization for Pareto Front Learning
• Controlled Text Generation as Continuous Optimization with Multiple Constraints
• Multi-task problems are not multi-objective
• Pareto Self-Supervised Training for Few-Shot Learning
• Adaptive Distillation: Aggregating Knowledge from Multiple Paths for Efficient Distillation
• Sequential Reptile: Inter-Task Gradient Alignment for Multilingual Learning
• Scalable Unidirectional Pareto Optimality for Multi-Task Learning with Constraints
• Balancing Average and Worst-case Accuracy in Multitask Learning