- Services & Solutions
- Clients & Partners
Advanced Research Computer Scientist, Speech Technology and Research Laboratory
Vikramjit Mitra, Ph.D., focuses on research involving signal processing, deep learning (with special focus on convolutional neural nets and deep autoencoders), robust speech applications (noise, reverberation, channel etc. robustness), automatic speech recognition, information retrieval, machine learning and auditory neuroscience.
At SRI, Mitra serves as principal or co-principal investigator (PI/co-PI) of National Science Foundation (NSF)-funded projects studying automatic transcription of low resource languages, production features for large vocabulary speech recognition and robust speech recognition. He has worked on several projects funded by the Defense Advanced Research Projects Agency (DARPA), Sandia National Laboratories, the Air Force Research Laboratory and other clients. In his work, he has proposed several robust acoustic features, signal processing techniques and deep learning architectures that have helped to produce state-of-the-art results in robust speaker recognition, speech recognition, language identification, speech activity detection and speech based mental health prediction (such as depression and suicidal state detection).
Mitra has authored more than 65 peer-reviewed papers in the areas of signal processing, robust acoustic features, speech recognition, language, and speaker recognition and has been an active reviewer of speech processing and machine learning conferences and journals. He is a senior member of the IEEE and an affiliate member of the Spoken Language Technology Committee (SLTC). He has served on NSF panels and on the scientific/technical committees of SPASR2013, MLSLP2012 and conferences/workshops such as ICASSP, Interspeech, ASRU and SLT.
Prior to joining SRI, Mitra worked as a research associate for the Speech Communication Laboratory at the University of Maryland, College Park, where he also received his Ph.D. in electrical engineering.