Hacking Task Confounder in Meta-Learning

TL;DR

This paper identifies task confounders in meta-learning that cause negative transfer, and proposes a causal representation method to improve generalization, achieving state-of-the-art results.

Contribution

It uncovers task confounders in meta-learning through causal analysis and introduces MetaCRL, a method to eliminate confounders and enhance meta-learning performance.

Findings

MetaCRL outperforms existing methods on benchmark datasets.

Task confounders cause negative transfer in meta-learning.

Causal analysis reveals spurious correlations affecting generalization.

Abstract

Meta-learning enables rapid generalization to new tasks by learning knowledge from various tasks. It is intuitively assumed that as the training progresses, a model will acquire richer knowledge, leading to better generalization performance. However, our experiments reveal an unexpected result: there is negative knowledge transfer between tasks, affecting generalization performance. To explain this phenomenon, we conduct Structural Causal Models (SCMs) for causal analysis. Our investigation uncovers the presence of spurious correlations between task-specific causal factors and labels in meta-learning. Furthermore, the confounding factors differ across different batches. We refer to these confounding factors as "Task Confounders". Based on these findings, we propose a plug-and-play Meta-learning Causal Representation Learner (MetaCRL) to eliminate task confounders. It encodes decoupled generating factors from multiple tasks and utilizes an invariant-based bi-level optimization mechanism to ensure their causality for meta-learning. Extensive experiments on various benchmark datasets demonstrate that our work achieves state-of-the-art (SOTA) performance.