Learning to Transfer: A Foliated Theory

Abstract

Learning to transfer considers learning solutions to tasks in a such way that relevant knowledge can be transferred from known task solutions to new, related tasks. This is important for general learning, as well as for improving the efficiency of the learning process. While techniques for learning to transfer have been studied experimentally, we still lack a foundational description of the problem that exposes what related tasks are, and how relationships between tasks can be exploited constructively. In this work, we introduce a framework using the differential geometric theory of foliations that provides such a foundation.

References in corpus (16)

• Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks
• Prototypical Networks for Few-shot Learning
• An Overview of Multi-Task Learning in Deep Neural Networks
• A continual learning survey: Defying forgetting in classification tasks
• Theoretical Models of Learning to Learn
• An Empirical Investigation of Catastrophic Forgetting in Gradient-Based Neural Networks
• Domain Adaptation for Statistical Classifiers
• Towards a Definition of Disentangled Representations
• An introduction to domain adaptation and transfer learning
• Fast Context Adaptation via Meta-Learning
• Online Learning: A Comprehensive Survey
• A short guide through integration theorems of generalized distributions
• On the Value of Target Data in Transfer Learning
• Disentangled Skill Embeddings for Reinforcement Learning
• Disentangling by Subspace Diffusion
• Continual Learning: Tackling Catastrophic Forgetting in Deep Neural Networks with Replay Processes