Dimitri M. Kanevsky

Dimitri M. Kanevsky is a computer scientist and engineer known for pioneering contributions to automated speech recognition, natural language processing, and computational linguistics. He worked at institutions including Bell Labs, Microsoft Research, and academic collaborations with Carnegie Mellon University, influencing systems used in telecommunications, military, and commercial technology applications.

Early life and education

Kanevsky was born in the Soviet Union and educated in institutions linked to Moscow State University, Lomonosov Moscow State University, and Soviet-era research establishments such as the Institute for Information Transmission Problems and the Soviet Academy of Sciences. He later emigrated and pursued graduate research connected to programs at Stanford University, Massachusetts Institute of Technology, and exchanges with researchers from Bell Labs and AT&T Laboratories. His formative mentors and collaborators included figures associated with Yuri G. Zhuravlev, Vladimir Lukin, and prominent Soviet computing scientists who interfaced with Western researchers at venues like the International Conference on Acoustics, Speech, and Signal Processing and the European Speech Communication Association.

Career

Kanevsky's professional career spans work at Soviet research centers, corporate laboratories, and academic appointments affiliated with Bell Labs, AT&T Bell Laboratories, Microsoft Research, and partnerships with Carnegie Mellon University and the University of California, Berkeley. He contributed to projects for Lucent Technologies, Nokia Research Center, and collaborations with the Defense Advanced Research Projects Agency and National Science Foundation-funded programs. He served on program committees for conferences such as ICASSP, ACL, and INTERSPEECH, and participated in editorial roles for journals like the IEEE Transactions on Audio, Speech, and Language Processing and the Computer Speech & Language journal.

Research and contributions

Kanevsky's research focused on statistical and algorithmic methods for automatic speech recognition, speaker adaptation, and phonetic modeling, building on frameworks developed at Bell Labs, CMU Sphinx, and within the Hidden Markov Model paradigm championed in work by researchers at Johns Hopkins University and IBM Research. He published on language modeling techniques related to the n-gram approach used by systems deployed by Google, Microsoft, and Yahoo! for voice search and dictation, and on noise-robust recognition methods referenced in studies at MIT Lincoln Laboratory and Honeywell. His work intersected with keyword spotting systems used in telephony platforms by AT&T and Verizon, and with multilingual speech technology projects involving European Commission initiatives and collaborations with the International Telecommunication Union.

Kanevsky contributed to algorithmic optimizations for decoder architectures utilized by projects at Dragon Systems, Philips Speech Processing, and open-source toolkits such as Kaldi (software) and HTK. He also engaged in research on corpus development and annotation practices comparable to efforts by Linguistic Data Consortium, Corpus of Contemporary American English, and the TIMIT Acoustic-Phonetic Continuous Speech Corpus. His interdisciplinary collaborations included work with researchers from University of Edinburgh, University of Cambridge, Max Planck Institute for Psycholinguistics, and applied research with Siemens and Bosch.

Awards and honors

Kanevsky received recognition from professional organizations including the Institute of Electrical and Electronics Engineers and the International Speech Communication Association. He was invited to deliver keynote or plenary addresses at conferences such as ICASSP, INTERSPEECH, and symposia organized by IEEE Signal Processing Society and the Association for Computational Linguistics. His contributions were acknowledged in workshops sponsored by the National Academy of Engineering and in panels convened by the Defense Advanced Research Projects Agency.

Selected publications

- Kanevsky, D. M., articles in proceedings of ICASSP and INTERSPEECH on robust speech recognition and language modeling techniques, coauthored with researchers from Carnegie Mellon University and Bell Labs. - Papers addressing decoder optimization and phonetic modeling appearing in IEEE Transactions on Audio, Speech, and Language Processing and Computer Speech & Language, with coauthors from Microsoft Research and AT&T Labs. - Contributions to corpus design and evaluation reported alongside work from the Linguistic Data Consortium and presented at ACL and European Conference on Speech Communication and Technology events.

Category:Computer scientists Category:Speech recognition researchers