A Foliated View of Transfer Learning

Abstract

Transfer learning considers a learning process where a new task is solved by transferring relevant knowledge from known solutions to related tasks. While this has been studied experimentally, there lacks a foundational description of the transfer learning problem that exposes what related tasks are, and how they can be exploited. In this work, we present a definition for relatedness between tasks and identify foliations as a mathematical framework to represent such relationships.

14 pages, 6 figures

References in corpus (3)

• An Overview of Multi-Task Learning in Deep Neural Networks
• What makes ImageNet good for transfer learning?
• Disentangled Skill Embeddings for Reinforcement Learning

Cited by in corpus (1)

• GENNI: Visualising the Geometry of Equivalences for Neural Network Identifiability