Can Active Learning Experience Be Transferred?

08/02/2016
by   Hong-Min Chu, et al.
0

Active learning is an important machine learning problem in reducing the human labeling effort. Current active learning strategies are designed from human knowledge, and are applied on each dataset in an immutable manner. In other words, experience about the usefulness of strategies cannot be updated and transferred to improve active learning on other datasets. This paper initiates a pioneering study on whether active learning experience can be transferred. We first propose a novel active learning model that linearly aggregates existing strategies. The linear weights can then be used to represent the active learning experience. We equip the model with the popular linear upper- confidence-bound (LinUCB) algorithm for contextual bandit to update the weights. Finally, we extend our model to transfer the experience across datasets with the technique of biased regularization. Empirical studies demonstrate that the learned experience not only is competitive with existing strategies on most single datasets, but also can be transferred across datasets to improve the performance on future learning tasks.

READ FULL TEXT

page 1

page 2

page 3

page 4

research
08/10/2014

Exponentiated Gradient Exploration for Active Learning

Active learning strategies respond to the costly labelling task in a sup...
research
06/25/2021

Multi-Domain Active Learning: A Comparative Study

Building classifiers on multiple domains is a practical problem in the r...
research
02/28/2023

Active Learning with Combinatorial Coverage

Active learning is a practical field of machine learning that automates ...
research
06/14/2023

Towards Balanced Active Learning for Multimodal Classification

Training multimodal networks requires a vast amount of data due to their...
research
08/07/2018

Barriers to Active Learning for Computer Science Faculty

Active learning is a proven pedagogical style that has demonstrated valu...
research
10/06/2015

Active Transfer Learning with Zero-Shot Priors: Reusing Past Datasets for Future Tasks

How can we reuse existing knowledge, in the form of available datasets, ...
research
01/24/2020

Active Learning for Entity Alignment

In this work, we propose a novel framework for the labeling of entity al...

Please sign up or login with your details

Forgot password? Click here to reset