Learning to Balance: Bayesian Meta-Learning for Imbalanced and Out-of-distribution Tasks

TL;DR

This paper introduces Bayesian TAML, a meta-learning model that adaptively balances meta-knowledge and task-specific learning, effectively handling imbalanced and out-of-distribution tasks by learning to decide the reliance on meta-knowledge.

Contribution

It proposes a Bayesian inference framework for adaptive balancing in meta-learning, addressing limitations of fixed assumptions in existing methods for imbalanced and diverse tasks.

Findings

Significantly outperforms existing meta-learning methods on imbalanced datasets.

Effective in handling out-of-distribution tasks through learned balancing.

Ablation studies confirm the importance of each component and the Bayesian approach.

Abstract

While tasks could come with varying the number of instances and classes in realistic settings, the existing meta-learning approaches for few-shot classification assume that the number of instances per task and class is fixed. Due to such restriction, they learn to equally utilize the meta-knowledge across all the tasks, even when the number of instances per task and class largely varies. Moreover, they do not consider distributional difference in unseen tasks, on which the meta-knowledge may have less usefulness depending on the task relatedness. To overcome these limitations, we propose a novel meta-learning model that adaptively balances the effect of the meta-learning and task-specific learning within each task. Through the learning of the balancing variables, we can decide whether to obtain a solution by relying on the meta-knowledge or task-specific learning. We formulate this objective into a Bayesian inference framework and tackle it using variational inference. We validate our Bayesian Task-Adaptive Meta-Learning (Bayesian TAML) on multiple realistic task- and class-imbalanced datasets, on which it significantly outperforms existing meta-learning approaches. Further ablation study confirms the effectiveness of each balancing component and the Bayesian learning framework.