Information Planning for Text Data

TL;DR

This paper demonstrates that information planning based on entropy and mutual information significantly accelerates learning in text data models, reducing the need for large training datasets.

Contribution

It introduces the application of entropy and mutual information-based planning to text data, improving learning efficiency for Naive Bayes, supervised LDA, and neural networks.

Findings

Planning outperforms random selection in text data learning.

Entropy and mutual information-based planning accelerates model training.

Applicable to various supervised models, including neural networks.

Abstract

Information planning enables faster learning with fewer training examples. It is particularly applicable when training examples are costly to obtain. This work examines the advantages of information planning for text data by focusing on three supervised models: Naive Bayes, supervised LDA and deep neural networks. We show that planning based on entropy and mutual information outperforms random selection baseline and therefore accelerates learning.