Preventing Hidden Information Leaks Using Author Attribution Methods and Neural Networks

Abstract

This paper addresses the detection of hidden information leakage through the use of text steganography. In this paper, we present research results on the study possibility to detect hidden leakages by detecting changes in users writing styles using neural networks and various types of text features. The framework for hidden leakages detection based on discovering changes in the author's writing style with deep neural networks (RNN, LSTM, GRU) was proposed. To evaluate the hidden leakages detection accuracy were carried out series of experiments on text corpus, contains Russian online texts. The experiments showed that the LSTM and character 4-grams allow achieving the accuracy of 87%. Text preprocessing significantly decreases accuracy.

Keywords

• hidden leakages
• text steganography
• information security