Sameer Khurana

smallest.ai · MIT PhD

I work on machine learning for spoken language technologies — speech recognition, speech translation, and self-supervised representation learning, with a focus on cross-lingual transfer. I completed my PhD at MIT CSAIL in James Glass's group.

I'm currently building conversational speech models at smallest.ai. Past research has spanned multilingual speech models (SAMU-XLSR, Magic Dust), unsupervised speech representation learning (ConvDMM, FDMM), audio classification (Whisper-AT, CMKD), and Arabic dialect identification.

email google scholar linkedin

Recent Work 2025 · preprints & in submission

Information bottleneck for multimodal alignment

Aligning Multimodal Representations through an Information Bottleneck

A. Almudévar, J. M. Hernández-Lobato, S. Khurana, R. Marxer, A. Ortega

ICML · 2025

Factorized RVQ-GAN For Disentangled Speech Tokenization

S. Khurana, D. Klement, A. Laurent, D. Bobos, J. Novosad, P. Gazdik, E. Zhang, Z. Huang, A. Hussein, R. Marxer, Y. Masuyama, R. Aihara, C. Hori, F. G. Germain, G. Wichern, J. Le Roux

Interspeech · 2025

HASRD framework for hierarchical disentangling

HASRD: Hierarchical Acoustic and Semantic Representation Disentanglement

A. Hussein, S. Khurana, G. Wichern, F. G. Germain, J. Le Roux

Interspeech · 2025

Continuous AR speech enhancement inference regimes

Investigating Continuous Autoregressive Generative Speech Enhancement

H. Yang, G. Wichern, R. Aihara, Y. Masuyama, S. Khurana, F. G. Germain, J. Le Roux

Interspeech · 2025

Late Fusion and Multi-Level Fission Amplify Cross-Modal Transfer in Text-Speech LMs

S. Cuervo, A. Moumen, Y. Labrak, S. Khurana, A. Laurent, M. Rouvier, P. Woodland, R. Marxer

Preprint · 2025

Selected Publications top 18 by citations · scholar →

DeepSol: A Deep Learning Framework for Sequence-Based Protein Solubility Prediction

S. Khurana, R. Rawi, K. Kunji, G.-Y. Chuang, H. Bensmail, R. Mall

Bioinformatics · 2018270 citations

Automatic Dialect Detection in Arabic Broadcast Speech

A. Ali, N. Dehak, P. Cardinal, S. Khurana, S. H. Yella, J. Glass, P. Bell, S. Renals

Interspeech · 2016178 citations

Whisper-AT noise robustness vs sound classification

Whisper-AT: Noise-Robust Automatic Speech Recognizers are Also Strong General Audio Event Taggers

Y. Gong, S. Khurana, L. Karlinsky, J. Glass

Interspeech · 2023149 citations

Robust Training of Vector Quantized Bottleneck Models

A. Łańcucki, J. Chorowski, G. Sanchez, R. Marxer, N. Chen, H. J. G. A. Dolfing, S. Khurana, T. Alumäe, A. Laurent

IJCNN · 2020107 citations

PARP: Prune, Adjust and Re-Prune for Self-Supervised Speech Recognition

CIJ. Lai, Y. Zhang, A. H. Liu, S. Chang, Y.-L. Liao, Y.-S. Chuang, K. Qian, S. Khurana, D. Cox, J. Glass

NeurIPS · 202199 citations

Unsupervised Domain Adaptation for Speech Recognition via Uncertainty Driven Self-Training

S. Khurana, N. Moritz, T. Hori, J. Le Roux

ICASSP · 202190 citations

SAMU-XLSR: Semantically-Aligned Multimodal Utterance-level Cross-Lingual Speech Representation

S. Khurana, A. Laurent, J. Glass

IEEE J. Selected Topics in Signal Processing · 202260 citations

QCRI Advanced Transcription System (QATS) for the Arabic Multi-Dialect Broadcast Media Recognition: MGB-2 Challenge

S. Khurana, A. Ali

IEEE SLT · 201657 citations

CMKD CNN/AST cross-model knowledge distillation

CMKD: CNN/Transformer-Based Cross-Model Knowledge Distillation for Audio Classification

Y. Gong, S. Khurana, A. Rouditchenko, J. Glass

IEEE TPAMI · 202553 citations

Exploiting Convolutional Neural Networks for Phonotactic Based Dialect Identification

M. Najafian, S. Khurana, S. Shon, A. Ali, J. Glass

ICASSP · 201852 citations

Convolutional Deep Markov Model graphical structure

A Convolutional Deep Markov Model for Unsupervised Speech Representation Learning

S. Khurana, A. Laurent, W.-N. Hsu, J. Chorowski, A. Łańcucki, R. Marxer, J. Glass

Interspeech · 202032 citations

NIIRF: Neural IIR Filter Field for HRTF Upsampling and Personalization

Y. Masuyama, G. Wichern, F. G. Germain, Z. Pan, S. Khurana, C. Hori, J. Le Roux

ICASSP · 202429 citations

Scenario-Aware Audio-Visual TF-GridNet for Target Speech Extraction

Z. Pan, G. Wichern, Y. Masuyama, F. G. Germain, S. Khurana, C. Hori, J. Le Roux

ASRU · 202327 citations

Whisper vs XLS-R hours per language family

Comparison of Multilingual Self-Supervised and Weakly-Supervised Speech Pre-Training for Adaptation to Unseen Languages

A. Rouditchenko, S. Khurana, S. Thomas, R. Feris, L. Karlinsky, H. Kuehne, D. Harwath, B. Kingsbury, J. Glass

Interspeech · 202326 citations

Transfer Learning Algorithm for cross-lingual wav2vec adaptation

Magic Dust for Cross-Lingual Adaptation of Monolingual wav2vec-2.0

S. Khurana, A. Laurent, J. Glass

ICASSP · 202226 citations

QMDIS CNN architecture for dialect identification

QMDIS: QCRI-MIT Advanced Dialect Identification System

S. Khurana, M. Najafian, A. Ali, T. Al Hanai, Y. Belinkov, J. R. Glass

Interspeech · 201725 citations

NeuroHeed+ joint speaker extraction and AAD

NeuroHeed+: Improving Neuro-steered Speaker Extraction with Joint Auditory Attention Detection

Z. Pan, G. Wichern, F. G. Germain, S. Khurana, J. Le Roux

ICASSP · 202421 citations

Factorial Deep Markov Model graphical models

A Factorial Deep Markov Model for Unsupervised Disentangled Representation Learning from Speech

S. Khurana, S. R. Joty, A. Ali, J. Glass

ICASSP · 201918 citations

Contact

The best way to reach me is by email at sameerkhurana10@gmail.com. I'm always happy to hear from students, collaborators, and people thinking about the same questions.