Incremental Active Learning for Optimal Generalization

Authors:
Masashi Sugiyama;Hidemitsu Ogawa
Affiliations:
Department of Computer Science, Tokyo Institute of Technology, Meguro-ku, Tokyo, 152-8552, Japan;Department of Computer Science, Tokyo Institute of Technology, Meguro-ku, Tokyo, 152-8552, Japan
Venue:
Neural Computation
Year:
2000

Citing 5
Cited 4

Neural networks and the bias/variance dilemma

Neural Computation
Information-based objective functions for active data selection

Neural Computation
Introduction

Machine Learning - Special issue on reinforcement learning
Neural network exploration using optimal experiment design

Neural Networks
Active learning with statistical models

Journal of Artificial Intelligence Research

The subspace information criterion for infinite dimensional hypothesis spaces

The Journal of Machine Learning Research
Compiled code simulation of analog and mixed-signal systems using piecewise linear modeling of nonlinear parameters: A case study for ΔΣ modulator simulation

Integration, the VLSI Journal
Pool-based active learning with optimal sampling distribution and its information geometrical interpretation

Neurocomputing
Improving importance estimation in pool-based batch active learning for approximate linear regression

Neural Networks

Quantified Score

Hi-index	0.00

Visualization

Abstract

The problem of designing input signals for optimal generalization is called active learning. In this article, we give a two-stage sampling scheme for reducing both the bias and variance, and based on this scheme, we propose two active learning methods. One is the multipoint search method applicable to arbitrary models. The effectiveness of this method is shown through computer simulations. The other is the optimal sampling method in trigonometric polynomial models. This method precisely specifies the optimal sampling locations.