A Sparse Plus Low-Rank Exponential Language Model for Limited Resource Scenarios

This paper describes a new exponential language model that decomposes the model parameters into one or more low-rank matrices that learn regularities in the training data and one or more sparse matrices that learn exceptions (e.g., keywords). The low-rank matrices induce continuous-space representat...

Full description

Saved in:
Bibliographic Details
Published in:IEEE/ACM transactions on audio, speech, and language processing Vol. 23; no. 3; pp. 494 - 504
Main Authors: Hutchinson, Brian, Ostendorf, Mari, Fazel, Maryam
Format: Journal Article
Language:English
Published: Piscataway IEEE 01.03.2015
The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
Subjects:
Adaptation models
Algorithms
Data models
exponential
History
Language model
log bilinear
low-hyphen
Mathematical models
Matrix decomposition
Natural language processing
Regularity
Representations
sparse
Sparse matrices
Speech
Training
ISSN:2329-9290, 2329-9304
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:This paper describes a new exponential language model that decomposes the model parameters into one or more low-rank matrices that learn regularities in the training data and one or more sparse matrices that learn exceptions (e.g., keywords). The low-rank matrices induce continuous-space representations of words and histories. The sparse matrices learn multi-word lexical items and topic/domain idiosyncrasies. This model generalizes the standard ℓ 1 -regularized exponential language model, and has an efficient accelerated first-order training algorithm. Language modeling experiments show that the approach is useful in scenarios with limited training data, including low resource languages and domain adaptation.
Bibliography:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
content type line 23
ISSN:2329-9290
2329-9304
DOI:10.1109/TASLP.2014.2379593