Towards Task Sampler Learning for Meta-Learning
This work addresses the problem of optimizing task sampling for meta-learning practitioners, offering an incremental improvement by refining sampling strategies rather than proposing a new paradigm.
This paper challenges the common belief that increasing task diversity always enhances meta-learning generalization, finding that no universal task sampling strategy guarantees optimal performance and that over-constraining diversity can cause under-fitting or over-fitting. It introduces Adaptive Sampler (ASr), a plug-and-play module that dynamically adjusts task weights based on diversity, entropy, and difficulty, showing clear advantages in experiments on benchmark datasets.
Meta-learning aims to learn general knowledge with diverse training tasks conducted from limited data, and then transfer it to new tasks. It is commonly believed that increasing task diversity will enhance the generalization ability of meta-learning models. However, this paper challenges this view through empirical and theoretical analysis. We obtain three conclusions: (i) there is no universal task sampling strategy that can guarantee the optimal performance of meta-learning models; (ii) over-constraining task diversity may incur the risk of under-fitting or over-fitting during training; and (iii) the generalization performance of meta-learning models are affected by task diversity, task entropy, and task difficulty. Based on this insight, we design a novel task sampler, called Adaptive Sampler (ASr). ASr is a plug-and-play module that can be integrated into any meta-learning framework. It dynamically adjusts task weights according to task diversity, task entropy, and task difficulty, thereby obtaining the optimal probability distribution for meta-training tasks. Finally, we conduct experiments on a series of benchmark datasets across various scenarios, and the results demonstrate that ASr has clear advantages.