Understanding what users like to do/need to get is critical in human computer interaction. When natural user interface like speech or natural language is used in human-computer interaction, such as in a spoken dialogue system or with an internet search engine, language understanding becomes an important issue. Intent understanding is about indentifying the action a user wants a computer to take or the information she/he would like to obtain, conveyed in a spoken utterance or a text query.
In this project, we develop robust data-driven technologies applicable todifferent domains, make them morepractical by leveraging large amount of unlabeled data via unsupervised/semi-supervised machine learning;by innovating machine learning algorithms that work better with less data or mismatched data; and by augmenting statistical models with domain knowledge obtainedin a semi-supervised fashion.Research activities fall into the following areas:
- Data-Driven Approaches to Spoken Language/Query Understanding
- Unsupervised/Semi-Supervised Learning
- Automatic/Semi-automatic Acquisition of Domain Knowledge
- Authoring Tools for Spoken Language Understanding
- Application of Intent Undrestanding Technology
We have contributed to Microsoft products from the following teams:
- Microsoft Live Search/Commerce Search
- Microsoft adCenter
- Microsoft Speech Component Group
- Tellme
- Xiao Li, Ye-Yi Wang, and Alex Acero, Extracting Structured Information from User Queries with Semi-Supervised Conditional Random Fields, in SIGIR, July 2009
- Young-In Song, Ye-Yi Wang, Yun-Cheng Ju, Mike Seltzer, Ivan Tashev, and Alex Acero, Voice Search of Structured Media Data, in International Conference on Acoustics, Speech and Signal Processing, Institute of Electrical and Electornic Engineers, Inc., Taipei, Taiwan, April 2009
- Ye-Yi Wang, Xiao Li, and Alex Acero, Inductive and Example-Based Learning for Text Classification, in Interspeech, International Speech Communication Association, Brisbane, Australia, September 2008
- Xiao Li, Ye-Yi Wang, and Alex Acero, Learning Query Intent from Regularized Click Graphs, in SIGIR'08: the 31st Annual ACM SIGIR conference on Research and Development in Information Retrieval, Association for Computing Machinery, Inc., Singapore, Singapore, July 2008
- Sibel Yaman, Li Deng, Dong Yu, Ye-Yi Wang, and Alex Acero, An integrative and discriminative technique for spoken utterance classification, in IEEE Trans. Audio, Speech, and Language Processing, vol. 16, no. 6, pp. 1207-1214, Institute of Electrical and Electronics Engineers, Inc., 2008
- Sibel Yaman, Li Deng, Dong Yu, Ye-Yi Wang, and Alex Acero, A Discriminative Training Framework using N-Best Speech Recognition Transcriptions and Scores for Spoken Utterance Classification, in Proc. of the International Conference on Acoustics, Speech and Signal Processing, Institute of Electrical and Electronics Engineers, Inc., Honolulu, Hawaii, U.S.A., April 2007
- Ye-Yi Wang and Alex Acero, Maximum Entropy Model Parameterization with Tf-Idf Weighted Vector Space Model, in IEEE Automatic Speech Recognition and Understanding Workshop, Institute of Electrical and Electronics Engineers, Inc., Kyoto, Japan, 2007
- Ye-Yi Wang, John Lee, and Alex Acero, Speech Utterance Classification Model Training without Manual Transcriptions, in IEEE International Conference on Acoustics, Speech and Signal Processing, Institute of Electrical and Electronics Engineers, Inc., Roulouse, France, 2006
- Ye-Yi Wang and Alex Acero, Rapid development of spoken language understanding grammars, in Speech Communication, vol. 48, no. 3-4, pp. 390-416, Elsevier , 2006
- Dong Yu, Yun-Cheng Ju, Ye-Yi Wang, and Alex Acero, N-Gram Based Filler Model for Robust Grammar Authoring, in International Conference on Acoustics, Speech, and Signal Processing., Institute of Electrical and Electronics Engineers, Inc., Toulouse, France, 2006
