Markov Processes for Enhanced Deepfake Generation and Detection

View on arXiv ← Back to list

Authors: Michael A. Kouritzin, Ian Zhang, Jyoti Bhadana, Seoyeon Park

Published: 2024-11-12 18:17:36+00:00

AI Summary

This paper introduces a Markov Observation Model (MOM) for deepfake generation and detection, comparing its performance to GAN, SVM, and BPF methods using coin flip data. MOM significantly outperforms other methods in both generation and, especially, detection of deepfakes.

Abstract

New and existing methods for generating, and especially detecting, deepfakes are investigated and compared on the simple problem of authenticating coin flip data. Importantly, an alternative approach to deepfake generation and detection, which uses a Markov Observation Model (MOM) is introduced and compared on detection ability to the traditional Generative Adversarial Network (GAN) approach as well as Support Vector Machine (SVM), Branching Particle Filtering (BPF) and human alternatives. MOM was also compared on generative and discrimination ability to GAN, filtering and humans (as SVM does not have generative ability). Humans are shown to perform the worst, followed in order by GAN, SVM, BPF and MOM, which was the best at the detection of deepfakes. Unsurprisingly, the order was maintained on the generation problem with removal of SVM as it does not have generation ability.

Key findings

The MOM achieved the best performance in deepfake detection, significantly outperforming GAN, SVM, and BPF. The MOM also demonstrated strong performance in deepfake generation. Human performance was the worst.

Approach

The authors propose a Markov Observation Model (MOM) for both generating and detecting deepfakes. The MOM uses a pairwise Markov chain, learning parameters via a Baum-Welch-like EM algorithm. Detection is performed using Bayesian model selection via Bayes factors.

Datasets

Synthetic coin flip data, including real sequences (generated randomly), human-generated fake sequences, and deepfakes generated using a Simulator algorithm, GANs, and the proposed MOM.

Model(s)

Generative Adversarial Networks (GANs), Support Vector Machines (SVMs), Branching Particle Filtering (BPF), Markov Observation Model (MOM).

Author countries

Canada, Canada, Canada, Canada

← Previous