Sciweavers

Free Online Productivity Tools i2Speak i2Symbol i2OCR iTex2Img iWeb2Print iWeb2Shot i2Type iPdf2Split iPdf2Merge i2Bopomofo i2Arabic i2Style i2Image i2PDF iLatex2Rtf Sci2ools

5

NAACL
1994

favoriteEmaildiscussreport

120views Computational Linguistics» more NAACL 1994»

A Hybrid Approach to Adaptive Statistical Language Modeling

13 years 5 months ago

A Hybrid Approach to Adaptive Statistical Language Modeling

Download acl.ldc.upenn.edu

We desert'be our latest attempt at adaptive language modeling. At the heart of our approachis a Maximum Entropy(ME) modelwhich inc.orlxnatesmanyknowledgesources in a consistentmanner. The othercomponentsareaselectiveunigramcache,aconditionalbigram cache,anda conventionalstatictrigram. Wedescribetheknowledge sourcesusedtobuildsucha modelwithARPA'sofficialWSJcorpus, and report on perplexity and word error rate results obtained with it. Then, threedifferentadaptationparadigmsare discussed,and an additional experiment, based on AP wiredata, is used to compare them.

Ronald Rosenfeld

Real-time Traffic

Adaptive Language Modeling | NAACL 1994 | NAACL 2007 | Othercomponentsareaselectiveunigramcache | Word Error Rate |

claim paper

Related Content

» TEGa hybrid approach to information extraction

» UB at CLEF2004 Cross Language Information Retrieval Using Statistical Language Models

» Developing a hybrid NP parser

» Multialign Combining Linguistic and Statistical Techniques to Improve Alignments for Adapt...

» Towards a Unified Approach to Memory and StatisticalBased Machine Translation

» A Hybrid MorphemeWord Representation for Machine Translation of Morphologically Rich Langu...

» A Hybrid Statistical Data Preprocessing Approach for LanguageIndependent Text Classificati...

» Hybrid semantic tagging for information extraction

» A Hybrid Shared Memory Execution Model for a Data Parallel Language with IO

Post Info
More Details (n/a)

Added	02 Nov 2010
Updated	02 Nov 2010
Type	Conference
Year	1994
Where	NAACL
Authors	Ronald Rosenfeld

Comments (0)