Hemlata Tak
Papers
1
Total Citations
8
H-Index
1
About
Hemlata Tak is a leading researcher in audio forensics and speech processing, with a primary focus on detecting AI-generated and deepfake speech. Her work bridges the critical gap between high-performing deep neural network classifiers and the need for model interpretability—a challenge central to building trust in automated forensic tools. In her highly cited 2025 study, "Investigating voiced and unvoiced regions of speech for audio deepfake detection," Tak systematically analyzes how different phonetic components influence detection accuracy, revealing that voiced regions often carry more discriminative cues for synthetic speech. This contribution not only advances the state-of-the-art in anti-spoofing but also provides a transparent, explainable framework for human evaluators to verify machine decisions. With over 8 citations in a short time, her research is already shaping next-generation audio authentication systems. Tak’s work is essential for students and researchers tackling the growing threat of voice deepfakes in security, media forensics, and digital trust.
Research Focus
Key Achievements
Top Papers
- 1