Andreas Stolcke

Andreas Stolcke is an American speech processing specialist.

Early life

Stolcke obtained his PhD in computer science in 1994 from University of California, Berkeley and at the same time was both doctoral research and a short-time postdoc at the International Computer Science Institute. Before working for Microsoft he was a senior research engineer at SRI International in Menlo Park, California. He was an associate editor of the IEEE Transactions on Audio Speech and Language Processing as well as co-author of Computer Speech and Language. He is a fellow of both ISCA and IEEE and is currently writes articles regarding language and speaker modeling as well as speech understanding.[1]

Research

In 1994 he used Hidden Markov models and stochastic context-free grammar to develop a Bayesian network which was a part of his dissertation.[2]

In 2002 he developed SRILM, a program based on C++ libraries, executables and helper scripts.[3] In September of the same year he discovered that by using prosodic model it is possible to predict whether an utterance is neutral or is annoying and/or frustrating to people.[4]

References

  1. ↑ "Andreas Stolcke". Retrieved 29 March 2017.
  2. ↑ Andreas Stolcke (July 1994). "Bayesian Learning of Probabilistic LanguageModels" (PDF). University of California, Berkeley.
  3. ↑ Andreas Stolcke (September 16, 2002). "SRILM-an extensible language modeling toolkit" (PDF). Interspeech. 2002: 2002.
  4. ↑ Jeremy Ang; Rajdip Dhillon; Ashley Krupski; Elizabeth Shriberg; Andreas Stolcke (September 2002). "Prosody-based automatic detection of annoyance and frustration in human-computer dialog" (PDF). Interspeech.
This article is issued from Wikipedia. The text is licensed under Creative Commons - Attribution - Sharealike. Additional terms may apply for the media files.