摘要: Traditional approaches to language modelling have relied on a fixed corpus of text inform the parameters probability distribution over word sequences. Increasing size often leads better-performing models, but no matter how large, is static entity, unable reflect information about events which postdate it. We introduce an online paradigm interleaves estimation and application model. present Bayesian approach modelling, in marginal probabilities trigram model are dynamically updated match topic being dictated system. also describe architecture prototype we implemented uses World Wide Web (WWW) as source information, provide results from some initial proof concept experiments.
sci-hub.se 参考文章(5)
Ronald Rosenfeld, Philip Clarkson, Statistical Language Modeling using the CMU-Cambridge Toolkit conference of the international speech communication association. ,(1997)
Kristie Seymore, Ronald Rosenfeld, Using Story Topics For Language Model Adaptation conference of the international speech communication association. ,(1997)
Richard Stephen Szeliski, Bayesian Modeling of Uncertainty in Low-Level Vision ,(2011)
K.-F. Lee, H.-W. Hon, R. Reddy, An overview of the SPHINX speech recognition system IEEE Transactions on Acoustics, Speech, and Signal Processing. ,vol. 38, pp. 600- 610 ,(1990) , 10.1109/29.45616
S. Della Pietra, V. Della Pietra, J. Lafferty, Inducing features of random fields IEEE Transactions on Pattern Analysis and Machine Intelligence. ,vol. 19, pp. 380- 393 ,(1997) , 10.1109/34.588021