Provable Meta-Learning of Linear Representations
This work addresses the lack of theoretical foundations in meta-learning, specifically for learning transferable representations in data-scarce regimes, offering provable guarantees for practitioners in machine learning.
The paper tackles the problem of meta-learning for multi-task linear regression by providing provably fast and sample-efficient algorithms to learn and transfer a common low-dimensional linear representation across tasks, complemented by information-theoretic lower bounds on sample complexity.
Meta-learning, or learning-to-learn, seeks to design algorithms that can utilize previous experience to rapidly learn new skills or adapt to new environments. Representation learning -- a key tool for performing meta-learning -- learns a data representation that can transfer knowledge across multiple tasks, which is essential in regimes where data is scarce. Despite a recent surge of interest in the practice of meta-learning, the theoretical underpinnings of meta-learning algorithms are lacking, especially in the context of learning transferable representations. In this paper, we focus on the problem of multi-task linear regression -- in which multiple linear regression models share a common, low-dimensional linear representation. Here, we provide provably fast, sample-efficient algorithms to address the dual challenges of (1) learning a common set of features from multiple, related tasks, and (2) transferring this knowledge to new, unseen tasks. Both are central to the general problem of meta-learning. Finally, we complement these results by providing information-theoretic lower bounds on the sample complexity of learning these linear features.