Paul A. Crook
SENIOR APPLIED SCIENTIST
- Statistical methods for Dialogue Management; especially POMDP based models.
- Statistical belief tracking of the state of dialogues.
- Automatic dialogue policy learning/optimization using Dynamic Programming and Reinforcement Learning.
- Spoken Dialogue Systems.
- General application of statistical Machine Learning (ML) techniques.
- Omar Zia Khan, Jean-Philippe Robichaud, Paul Crook, and Ruhi Sarikaya, Hypotheses Ranking and State Tracking for a Multi-Domain Dialog System using Multiple ASR Alternates, in Proceedings of the 16th Annual Conference of the International Speech Communication Association (INTERSPEECH 2015), ISCA - International Speech Communication Association, September 2015.
- Yi Ma, Paul A. Crook, Ruhi Sarikaya, and Eric Fosler-Lussier, KNOWLEDGE GRAPH INFERENCE FOR SPOKEN DIALOG SYSTEMS, in Proceedings of 40th IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) 2015, IEEE – Institute of Electrical and Electronics Engineers, 19 April 2015.
- Jean-Philippe Robichaud, Paul A. Crook, Puyang Xu, Omar Zia Khan, and Ruhi Sarikaya, Hypotheses Ranking for Robust Domain Classification And Tracking in Dialogue Systems, in Proceedings of the 15th Annual Conference of the International Speech Communication Association (INTERSPEECH 2014), ISCA - International Speech Communication Association, September 2014.
- P.A. Crook, S. Keizer, Z. Wang, W. Tang, and O. Lemon, Real user evaluation of a POMDP spoken dialogue system using automatic belief compression, in Computer Speech and Language, vol. 28, no. 4, pp. 873-887, Elsevier, May 2014.