Many thanks to...
The speech group at the Cambridge Research Lab, especially Pedro Moreno for his Matlab help.
E. Levin and R. Pieraccini, “A stochastic model of computer-human interaction for learning dialog strategies,” in Proc. EUROSPEECH’97.
E. Levin, R. Pieraccini, and W. Eckert, “Using Markov decision processes for learning dialog strategies,” in Proc. ICASSP’98.
M. Walker, J. Fromer, and S. Narayanan, “Learning optimal dialog strategies: A case study of a spoken dialog agent for email,” in Proc. ACL’98.