Purvi Agrawal

Ph.D. Scholar @ LEAP Lab, Department of Electrical Engineering, Indian Institute of Science

Thesis

P. Agrawal, " Neural Representation Learning for Speech and Audio Signals", Indian Institute of Science, 2020. [pdf]

PhD Thesis Defense Talk

P. Agrawal, Neural Representation Learning of Speech and Audio Signals, Thesis Defense Talk, January 2021. [Video]

Colloquium

P. Agrawal, Neural Representation Learning of Speech and Audio Signals, Thesis Colloquium Talk, July 2020. [Video]

Publications

Journals

• P. Agrawal, S. Ganapathy, "Interpretable Representation Learning for Speech and Audio Signals Based on Relevance Weighting," IEEE Transactions and Audio, Speech and Language Processing, 2020.

• P. Agrawal, S. Ganapathy, "Modulation Filter Learning Using Deep Variational Networks for Robust Speech Recognition", IEEE Journal of Selected Topics in Signal Processing (J-STSP), Special Issue on Data Science: Machine Learning for Audio Signal Processing, April 2019.

• P. Agrawal and S. Ganapathy, "Unsupervised Modulation Filter Learning for Noise-Robust Speech Recognition", Journal of Acoustical Society of America, Sept. 2017.

Conferences

• P. Agrawal and S. Ganapathy,"Representation Learning For Speech Recognition Using Feedback Based Relevance Weighting", ICASSP, 2021. [arXiv]

• P. Agrawal and S. Ganapathy,"Robust Raw Waveform Speech Recognition Using Relevance Weighted Representations", INTERSPEECH, 2020.

• P. Agrawal and S. Ganapathy, "Unsupervised Raw Waveform Representation Learning for ASR", INTERSPEECH, 2019.

• P. Agrawal, S. Ganapathy, "Deep variational filter learning models for speech recognition", ICASSP, 2019.

• P. Agrawal, S. Ganapathy, "Comparison of unsupervised modulation filter learning methods for ASR", INTERSPEECH, 2018.

• N. Takahashi, P. Agrawal, N. Goswami and Y. Mitsufuji, "PhaseNet: Discretized phase modeling with deep neural networks for audio source separation", INTERSPEECH, 2018.

• P. Agrawal and S. Ganapathy "Speech representation learning using unsupervised data-driven modulation filtering for robust ASR", INTERSPEECH, Aug. 2017.

Challenges

• Low Resource Speech Recognition Challenge for Indian Languages - Finished 3rd on leaderboard with team - Purvi Agrawal, Sonali Singh, Jayanth Shankar - IITB, Dr. Sriram Ganapathy, Dr. Preethi Jyothi - IITB.


• The 5th CHiME Speech Separation and Recognition Challenge - Finished in top 10 teams for single-array track with team - Dr. Sriram Ganapathy, Purvi Agrawal.