PPL / Dataset Treebank / Treebank / Penn Treebank / Model M. Dataset Treebank / / /
IndustryTerm
neural network / recurrent neural networks / neural information processing systems / recursive binary search procedure / recurrent neural network / dynamic programming algorithm / quadratic time solution / computing / local search procedure / /
Organization
Association of Computational Linguistics / /
Person
Stefan Kombrink / Holger Schwenk / P.F. Brown / V / Sanjeev Khudanpur / Lukas Burget / Daniel Povey / Tomas Mikolov / J. Della Pietra / Jan Cernocky / Hai-Son Le / Dan Povey / Jrg Liermann / Anoop Deoras / Sven Martin / Martin Karafiat / Hermann Ney / Konstantin Makarychev / /
Position
author / RT / RNN-RT / /
PublishedMedium
Computational Linguistics / /
Technology
speech recognition / neural network / end end end Algorithm / dynamic programming algorithm / Machine Learning / previous classing algorithms / DP algorithm / exchange algorithm / /
SPEED REGULARIZATION AND OPTIMALITY IN WORD CLASSING Geoffrey Zweig and Konstantin Makarychev Microsoft Research {gzweig,komakary}@microsoft.com ABSTRACT Word-classing has been used in language modeling for two