I am Dr. Souvik Kundu

Natural Language Processing and Deep Learning Researcher

Souvik Kundu

Data and Applied Scientist, Microsoft, USA

Email: sokun[at]microsoft[dot]com

Research Interest

Applied Machine Learning, Deep Learning

Text-to-SQL, Semantic Parsing, Question Answering, Multi-step Reasoning, Dialogue Systems, Information Extraction, Natural Language Understanding

Automatic Speech Recognition, Acoustic Factor Adaptation, Noise Robustness, Model-based Adaptation

About me

I am a Data and Applied Scientist at Microsoft. I did my PhD at the department of Computer Science, National University of Singapore, under the supervision of Prof Hwee Tou Ng. My research interests are primarily in Applied Machine Learning and Deep Learning. My PhD thesis is focused on Neural Network-based Question Answering (QA) systems. In the starting years of my PhD, I also worked with Dr. Khe Chai Sim (now in Google).


2014-2020: Doctor of Philosophy (PhD), Computer Science

2010-2014: Bachelor of Engineering, Electronics and Telecommunication


2019-Present: Microsoft, Redmond, USA.

2018: Research Intern, Allen Institute for Artificial Intelligence, Seattle, USA.

2015: Student Research Intern, Jelinek Summer Workshop of Speech and Language Technology, University of Washington, Seattle, USA.


Question Answering Using Deep Neural Networks: Single Turn and Beyond
Souvik Kundu,
2020, PhD Dissertation

Learning to Identify Follow-up Questions in Conversational Question Answering
[pdf][code & data]
Souvik Kundu, Qian Lin, Hwee Tou Ng
ACL 2020

Exploiting Explicit Paths for Multi-hop Reading Comprehension
Souvik Kundu, Tushar Khot, Ashish Sabharwal, Peter Clark
ACL 2019

A Nil-Aware Answer Extraction Framework for Question Answering
Souvik Kundu, Hwee Tou Ng
EMNLP 2018

A Question-Focused Multi-Factor Attention Network for Question Answering
Souvik Kundu, Hwee Tou Ng
AAAI 2018

Adaptation of Deep Neural Network Acoustic Models for Robust Automatic Speech Recognition
[book chapter]
Khe Chai Sim, Yanmin Qian, Gautam Mantena, Lahiru Samarakoon, Souvik Kundu, Tian Tan
New Era for Robust Speech Recognition 2017 (by Springer Link)

Incorporating a Generative Front-End Layer to Deep Neural Network for Noise Robust Automatic Speech Recognition
Souvik Kundu, Khe Chai Sim, Mark Gales

Joint Acoustic Factor Learning for Robust Deep Neural Network Based Automatic Speech Recognition
Souvik Kundu, Gautam Mantena, Yanmin Qian, Tian Tan, Marc Delcroix, Khe Chai Sim

Speaker-Aware Training of LSTM-RNNs for Acoustic Modelling
Tian Tan, Yanmin Qian, Dong Yu, Souvik Kundu, Liang Lu, Khe Chai Sim, Xiong Xiao, Yu Zhang

See the full publication list in my Google Scholar.


Teaching Experience

Spring 2015: Software Testing (CS4218)

Fall 2015: Data Structures and Algorithms I (CS1020E)

Spring 2016: Software Testing (CS4218)

Fall 2016: Programming Methodology (CS1101S)

Spring 2017: Data Structures and Algorithms I (CS1020 & CS1020E)