LibMOON: A Gradient-based MultiObjective OptimizatioN Library in PyTorch
This addresses the problem of inefficient scaling in MOPs for researchers and practitioners, though it is incremental as it builds on existing gradient-based techniques.
The paper tackles the lack of scalable gradient-based methods for multiobjective optimization problems (MOPs) in machine learning by introducing LibMOON, the first library supporting state-of-the-art gradient-based methods, which enables optimization for large-scale models with thousands or millions of parameters.
Multiobjective optimization problems (MOPs) are prevalent in machine learning, with applications in multi-task learning, learning under fairness or robustness constraints, etc. Instead of reducing multiple objective functions into a scalar objective, MOPs aim to optimize for the so-called Pareto optimality or Pareto set learning, which involves optimizing more than one objective function simultaneously, over models with thousands / millions of parameters. Existing benchmark libraries for MOPs mainly focus on evolutionary algorithms, most of which are zeroth-order / meta-heuristic methods that do not effectively utilize higher-order information from objectives and cannot scale to large-scale models with thousands / millions of parameters. In light of the above gap, this paper introduces LibMOON, the first multiobjective optimization library that supports state-of-the-art gradient-based methods, provides a fair benchmark, and is open-sourced for the community.