Daniyal Kabir Dar

iProbe Lab, Dept. of Computer Science and Engineering

Michigan State University

East Lansing, MI

I’m a PhD student in Computer Science at Michigan State University, where I work in the iProbe Lab advised by Dr. Arun Ross. My research sits at the intersection of audio deepfake detection, speaker identity bias, adversarial robustness in speech systems, and explainability in biometric systems.

Broadly, I’m interested in a question that keeps coming up in my work: when a detector says “fake” or “real,” how much of that decision is actually about synthesis artifacts, and how much is it quietly leaning on who is speaking? Most of my projects — from measuring identity sensitivity in deepfake detectors to studying how adversarial perturbations shift speaker identity at the phonetic level — try to pull that question apart and build tools that make these systems more reliable and more explainable.

I have papers accepted at ICASSP 2026 (oral) and IJCB 2026, with a third currently under review at SLT 2026. My current project builds a contrastive, speaker-normalized system for generating acoustic explanations of deepfake detector decisions.

Outside of research, I spend time thinking about applied and entrepreneurial ideas in voice AI and agentic tools, and I game on a dual-GPU setup when I get the chance.

Feel free to reach out via email if you’d like to chat about audio forensics, biometrics, or speech ML.

news

Jul 23, 2026	New preprint on arXiv: “Probing Speaker Identity Sensitivity in Audio Deepfake Detectors,” accepted at IJCB 2026.
May 01, 2026	Presented “Impact of Phonetics on Speaker Identity in Adversarial Voice Attack” as an oral presentation at ICASSP 2026 in Barcelona.
Jan 01, 2026	Our paper “Impact of Phonetics on Speaker Identity in Adversarial Voice Attack” was accepted for an oral presentation at ICASSP 2026.

selected publications

IJCB
Probing Speaker Identity Sensitivity in Audio Deepfake Detectors

Daniyal Kabir Dar and Arun Ross

IEEE/IAPR International Joint Conference on Biometrics (IJCB), 2026

Abs Bib HTML

We propose the Identity Sensitivity Score (ISS), a per-utterance diagnostic that measures how much an audio deepfake detector’s output changes across different speaker-identity contexts, without requiring ground-truth labels at inference time. Across two detectors and two datasets, misclassified utterances show ISS scores tens of times higher than correctly classified ones, indicating that many detector errors are tied to speaker-identity reliance rather than synthesis artifacts alone.
@article{dar2026probing, title = {Probing Speaker Identity Sensitivity in Audio Deepfake Detectors}, author = {Dar, Daniyal Kabir and Ross, Arun}, journal = {IEEE/IAPR International Joint Conference on Biometrics (IJCB)}, year = {2026}, archiveprefix = {arXiv}, primaryclass = {cs.SD}, }
ICASSP
Impact of Phonetics on Speaker Identity in Adversarial Voice Attack

Daniyal Kabir Dar, Qiben Yan, Li Xiao, and 1 more author

In IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2026

Abs Bib HTML

We analyze adversarial audio perturbations at the phonetic level, showing that they exploit systematic confusions such as vowel centralization and consonant substitutions. These distortions degrade phonetic cues that are critical for speaker verification, leading to measurable identity drift under adversarial attack.
@inproceedings{dar2026impact, title = {Impact of Phonetics on Speaker Identity in Adversarial Voice Attack}, author = {Dar, Daniyal Kabir and Yan, Qiben and Xiao, Li and Ross, Arun}, booktitle = {IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)}, pages = {13462--13466}, year = {2026}, publisher = {IEEE}, archiveprefix = {arXiv}, primaryclass = {cs.SD}, }