Scalable Unidirectional Pareto Optimality for Multi-Task Learning with Constraints
This work addresses the challenge of balancing competing objectives under constraints in multi-task learning, offering a scalable solution for applications like image classification.
The paper tackles the problem of multi-objective optimization in multi-task learning by introducing a method to learn the full Pareto manifold at train-time, enabling users to select any optimal trade-off at run-time. Results show consistent improvements in accuracy and efficiency over prior methods on synthetic benchmarks and image classification tasks.
Multi-objective optimization (MOO) problems require balancing competing objectives, often under constraints. The Pareto optimal solution set defines all possible optimal trade-offs over such objectives. In this work, we present a novel method for Pareto-front learning: inducing the full Pareto manifold at train-time so users can pick any desired optimal trade-off point at run-time. Our key insight is to exploit Fritz-John Conditions for a novel guided double gradient descent strategy. Evaluation on synthetic benchmark problems allows us to vary MOO problem difficulty in controlled fashion and measure accuracy vs. known analytic solutions. We further test scalability and generalization in learning optimal neural model parameterizations for Multi-Task Learning (MTL) on image classification. Results show consistent improvement in accuracy and efficiency over prior MTL methods as well as techniques from operations research.