STEGanography via Arithmetic coding and Strong neURAl modelS

This repository contains implementations of the steganography algorithms from "Neural Linguistic Steganography," Zachary M. Ziegler*, Yuntian Deng*, Alexander M. Rush.

There is a breaking change in pytorch 1.2, make sure to use pytorch 1.0 as in requirements.txt.

Our online demo can be found at https://steganography.live/.

Experiments in the paper use the medium (345M parameter) GPT model via pytorch_transformers. For compute reasons the default in this code base is the small version but the medium or large versions can be used by changing the model_name parameter of get_model.

The steganography algorithms implemented are:

1. Our proposed arithmetic coding-based algorithm
2. The Huffman algorithm from RNN-Stega: Linguistic Steganography Based on Recurrent Neural Networks
3. The binning algorithm from Generating Steganographic Text with LSTMs

An example of encoding and decoding a message is in run_single.py. The algorithm used is determined by the mode parameter.