Year: 2019
Type: preprint
Abstract: We build a theoretical framework for designing and understanding practical meta-learning methods that integrates sophisticated formalizations of task-similarity with the extensive literature on online convex optimization and sequential prediction algorithms. Our approach enables the task-similarity to be learned adaptively, provides sharper transfer-risk bounds in the setting of statistical learning-to-learn, and leads to straightforward derivations of average-case regret bounds for efficient algorithms in settings where the task-environment changes dynamically or the tasks share a certain geometric structure. We use our theory to modify several popular meta-learning algorithms and improve their meta-test-time performance on standard problems in few-shot learning and federated learning.
Source: arXiv (Cornell University)
Institution: Carnegie Mellon University
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
Awards: AitF: FULL: From Worst-Case to Realistic-Case Analysis for Large Scale Machine Learning Algorithms, BIGDATA: F: Optimization in Federated Networks of Devices, III: Medium: Scalable Machine Learning for Genome-Wide Association Analyses, RI: AF: Small: Collaborative Research: Differentially Private Learning: From Theory To Applications, AF: Small: Learning Theory for a Modern World: Transfer Learning, Unsupervised Learning, and Beyond Prediction, RI: Medium: Learning to Search: Provable Guarantees and Applications
OpenAlex ID: W2948168221