Anup Singh

Computer Science Ph.D. Student at Ghent University

anup_final.png

I am currently a Ph.D. student in the Department of Electronics and Information Systems at Ghent University, where I am a member of the Speech Group at IDLab. I am advised by Prof. Kris Demuynck and Prof. Vipul Arora. As part of my PhD, I also spent time as a long-term research visitor at MADHAV Lab, IIT Kanpur.

My research focuses on self-supervised learning for speech and audio processing, with an emphasis on robust and efficient audio indexing and retrieval. More recently, I have been exploring speech tokenization, with the aim of contributing to emerging directions such as textless NLP and speech-LLMs.

I hold a BS–MS dual degree in Mathematics from the Indian Institute of Science Education and Research (IISER-Kolkata). During my undergraduate studies, I worked on various machine-learning-oriented projects, which shaped my interest in applied machine learning.

Outside of work, I enjoy learning about geopolitics, reading, and playing sports (mostly lawn-tennis these days!)

news

Jan 22, 2026 Our paper titled “Harmonic Summation-Based Robust Pitch Estimation in Noisy and Reverberant Environments” has been accepted at NCC 2026. Check out the paper.
Jan 17, 2026 Our paper titled “BEST-STD2.0: Balanced and Efficient Speech Tokenizer for Spoken Term Detection” has been accepted at ICASSP 2026. Check out the paper.
Jan 01, 2026 I have joined Amazon as an Applied Scientist II, working on Speech LLMs.

latest posts

Feb 21, 2026 Flow Matching
Dec 26, 2025 What are Diffusion Models?
Dec 18, 2025 Variational AutoEncoders

selected publications

  1. Interspeech
    interspeech2025.png
    Language-Agnostic Speech Tokenizer for Spoken Term Detection with Efficient Retrieval
    Anup Singh, Kris Demuynck, and Vipul Arora
    In Proc. Interspeech 2025, 2025
  2. ICASSP
    icassp2025.png
    BEST-STD: Bidirectional Mamba-Enhanced Speech Tokenization for Spoken Term Detection
    Anup Singh, Kris Demuynck, and Vipul Arora
    In ICASSP 2025-2025 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2025