A Mutation-based Text Generation for Adversarial Machine Learning Applications

TL;DR

This paper introduces mutation-based methods for generating text in adversarial machine learning, emphasizing the use of human text samples and exploring potential extensions for improved application-specific mutation operators.

Contribution

It presents a novel mutation-based approach for text generation in adversarial contexts, differing from traditional machine-generated methods by requiring human input samples.

Findings

Mutation operators can produce diverse adversarial texts

Mutation-based generation requires human text samples

Potential for extending mutation operators based on application needs

Abstract

Many natural language related applications involve text generation, created by humans or machines. While in many of those applications machines support humans, yet in few others, (e.g. adversarial machine learning, social bots and trolls) machines try to impersonate humans. In this scope, we proposed and evaluated several mutation-based text generation approaches. Unlike machine-based generated text, mutation-based generated text needs human text samples as inputs. We showed examples of mutation operators but this work can be extended in many aspects such as proposing new text-based mutation operators based on the nature of the application.