Learning speech semantics with keyword classification trees
Roland Kühn, Renato De Mori · IEEE International Conference on Acoustics Speech and Signal Processing · 1993
A linguistic analyzer based on KCTs (keyword classification trees) was trained on sentences from the ATIS (Air Travel Information System) air travel task and incorporated into the system (CHANEL) built at CRIM (Centre de Recherche Informatique de Montreal) for the Nov. 1992 ATIS benchmarks. Word sequences were processed by a local parser that identified semantically important noun phrases and then passed through a forest of KCTs, each responsible for generating a different aspect of the semantic representation. CHANEL attained a reasonable performance level, despite its heavy reliance on KCTs rather than on handcoded linguistic rules. The CRIM speech recognition system had a recognition rate of 88.9% words correct; CHANEL is clearly robust.>