My Projects

DeepFake-o-meter

An open-source platform for robust deepfake detection platform, supporting various media types including audio and video. Continuously updated with the latest detection algorithms.

Visit Project Website

Roles: Project Manager

Forensic Analysis of Segmental Speech Features, 2025

Investigating the use of segmental speech features for more robust and nuanced deepfake audio detection, building on collaboration with other researchers.

View Related Preprint

Roles: Co-Researcher

Landscape Survey of Generative AI Technologies and Their Real-World Impacts, NSF Center for Identification Technology Research (CITeR), 2024 - 2026, $400,000.

We aim to survey the landscape of DeepFake generation methods and their state-of-the-art performance.

Visit Project Website

Roles: Research Assistant

NSF Convergence Accelerator: Track F: Deception Awareness and Resilience Training (DART), National Science Foundation, Project 2137871, 2022-2024, $5,000,000

DART is a cross-disciplinary, multi-institutional, user-centered effort to develop games to help older adults recognize and navigate scams.

Visit Project Website

Roles: Research Assistant

AI Synthetic Video Detection Using Multi-Model Manipulation, 24F-02B (NSF CITeR) $50,000

Our objective is to develop a multi-modal deepfake detection model that integrates audio and video analysis to overcome the limitations of existing methods.

Roles: Co-Researcher

A Comprehensive Dataset for Deepfake Audio Detection, 24S-01B (NSF CITeR) $50,000

Development and curation of a large-scale, diverse dataset specifically designed to train and benchmark deepfake audio detection models.

Roles: Lead Researcher

A benchmark dataset for neural vocoder identificatio 22S-01B (NSF CITeR) $50,000

Research project focusing on identifying unique artifacts left by neural vocoders in AI-synthesized voices to improve deepfake audio detection accuracy.

View Related Publication

Roles: Lead Researcher

INSuRE+C 2021 Audio Deepfake Detection

Our objective is to develop methodologies for synthetic audio detection through the identification of the neural vocoder employed during the generation process.

Visit Project Website

Roles: Lead Researcher