Assisting Transfer-Enabled Machine Learning Algorithms: Leveraging Human Knowledge for Curriculum Design

Abstract

Transfer learning is a successful technique that significantly improves machine learning algorithms by training on a sequence of tasks rather than a single task in isolation. However, there is currently no systematic method for deciding how to construct such a sequence of tasks. In this paper, I propose that while humans are well-suited for the task of curriculum development, significant research is still necessary to better understand how to create effective curricula for machine learning algorithms. Transfer Learning Traditional machine learning algorithms often require a large amount of data to solve a given task, even when similar tasks have already been solved. The insight of transfer learning (TL) is to make use of data from one or more previous tasks in order to learn the current task with less data, which may be expensive to gather. Generalization is possible not only within tasks, but also across tasks. This insight is not new; transfer has been studied in the psychological literature (Thorndike & Woodworth 1901; Skinner 1953) for many years. Transfer learning has been gaining in popularity in recent years as researchers have successfully applied TL techniques to classification tasks (Thrun 1996;