Harshvardhan Takawale

Brendan Iribe Center, Room 3245

University of Maryland, College Park

I am a fourth-year Computer Science PhD candidate at the University of Maryland, College Park, advised by Prof. Nirupam Roy in the iCoSMoS Lab.

My research builds physics-informed machine learning systems for spatial acoustics, RF sensing, and ambient computing — enabling rich perception and inference on wearable and low-power platforms embedded in everyday objects and environments. I am interested in how resource-constrained, intermittently-powered systems can perceive gestures, micro-motions, directions, and physical context for environment-sensing and ubiquitous-healthcare applications.

I have been fortunate to intern at Dolby Laboratories (San Francisco), where I worked on modelling car-cabin acoustic frequency-response fields with implicit neural representations, and at Nokia Bell Labs (Cambridge, UK) in the Pervasive Systems group on detecting auditory attention from in-ear muscle contractions on earables. Earlier, I was a Lead Researcher at Silence Laboratories (Singapore). I received my B.E. from BITS Pilani in 2020.

Research interests: Spatial Acoustics · Physics-informed Machine Learning · Multimodal Imaging · RF Sensing · Ambient Computing · Ubiquitous Healthcare · Low-power Learning Systems

Feel free to reach out by email — I’m always happy to chat about research, collaborations, or internships.

news

May 15, 2026	I will be returning to Dolby Laboratories (San Francisco) as a Research Intern in Summer 2026.
May 01, 2026	Selected for the Financial Aid Award and the Gold Reviewer Award at ICML 2026.
Apr 01, 2026	Our paper INFER: Learning Implicit Neural Frequency Response Fields for Confined Car Cabin is accepted at ICML 2026! 🎉
Mar 01, 2026	Selected as a Principal Investigator for the Seed Award Program by the Artificial Intelligence Interdisciplinary Institute at Maryland (AIM).
Feb 01, 2026	Selected for the NSF I-Corps program as Entrepreneurial Lead.
Dec 15, 2025	I cleared my PhD Preliminary Exam and am now a PhD Candidate!

selected publications

INFER: Learning Implicit Neural Frequency Response Fields for Confined Car Cabin

Harshvardhan Takawale, Nirupam Roy, and Phil Brown

In International Conference on Machine Learning (ICML), 2026

Abs arXiv Bib HTML

Accurate modeling of spatial acoustics is critical for immersive and intelligible audio in confined, resonant environments such as car cabins. We propose INFER (Implicit Neural Frequency Response fields), a frequency-domain neural framework jointly conditioned on source and receiver positions and orientations to directly learn complex-valued frequency response fields inside confined, resonant environments. Our approach significantly outperforms time- and hybrid-domain baselines on both simulated and real-world automotive datasets, cutting average magnitude and phase reconstruction errors by over 39% and 51%, respectively.
@inproceedings{takawale2026infer, title = {INFER: Learning Implicit Neural Frequency Response Fields for Confined Car Cabin}, author = {Takawale, Harshvardhan and Roy, Nirupam and Brown, Phil}, booktitle = {International Conference on Machine Learning (ICML)}, year = {2026}, }

Towards Detecting Auditory Attention from in-Ear Muscle Contractions using Commodity Earbuds

Harshvardhan Takawale, Yang Liu, Khaldoon Al-Naimi, and 2 more authors

In IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2025

Bib PDF

@inproceedings{takawale2025auditory,
  title = {Towards Detecting Auditory Attention from in-Ear Muscle Contractions using Commodity Earbuds},
  author = {Takawale, Harshvardhan and Liu, Yang and Al-Naimi, Khaldoon and Kawsar, Fahim and Montanari, Alessandro},
  booktitle = {IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)},
  year = {2025},
}

Learning Speaker-Listener Mutual Head Orientation by Leveraging HRTF and Voice Directivity on Headphones

Harshvardhan Takawale and Nirupam Roy

In IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2024

Bib PDF

@inproceedings{takawale2024headorientation,
  title = {Learning Speaker-Listener Mutual Head Orientation by Leveraging HRTF and Voice Directivity on Headphones},
  author = {Takawale, Harshvardhan and Roy, Nirupam},
  booktitle = {IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)},
  year = {2024},
}

Scribe: Simultaneous Voice and Handwriting Interface

Yang Bai, Irtaza Shahid, Harshvardhan Takawale, and 1 more author

Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies (IMWUT), 2024

Bib PDF Slides

@article{bai2024scribe,
  title = {Scribe: Simultaneous Voice and Handwriting Interface},
  author = {Bai, Yang and Shahid, Irtaza and Takawale, Harshvardhan and Roy, Nirupam},
  journal = {Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies (IMWUT)},
  year = {2024},
}