A Fast Variational Approach for Learning Markov Random Field Language Models

Yacine Jernite, Alexander Rush, David Sontag
Proceedings of the 32nd International Conference on Machine Learning, PMLR 37:2209-2217, 2015.

Abstract

Language modelling is a fundamental building block of natural language processing. However, in practice the size of the vocabulary limits the distributions applicable for this task: specifically, one has to either resort to local optimization methods, such as those used in neural language models, or work with heavily constrained distributions. In this work, we take a step towards overcoming these difficulties. We present a method for global-likelihood optimization of a Markov random field language model exploiting long-range contexts in time independent of the corpus size. We take a variational approach to optimizing the likelihood and exploit underlying symmetries to greatly simplify learning. We demonstrate the efficiency of this method both for language modelling and for part-of-speech tagging.

Cite this Paper


BibTeX
@InProceedings{pmlr-v37-jernite15, title = {A Fast Variational Approach for Learning Markov Random Field Language Models}, author = {Jernite, Yacine and Rush, Alexander and Sontag, David}, booktitle = {Proceedings of the 32nd International Conference on Machine Learning}, pages = {2209--2217}, year = {2015}, editor = {Bach, Francis and Blei, David}, volume = {37}, series = {Proceedings of Machine Learning Research}, address = {Lille, France}, month = {07--09 Jul}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v37/jernite15.pdf}, url = {https://proceedings.mlr.press/v37/jernite15.html}, abstract = {Language modelling is a fundamental building block of natural language processing. However, in practice the size of the vocabulary limits the distributions applicable for this task: specifically, one has to either resort to local optimization methods, such as those used in neural language models, or work with heavily constrained distributions. In this work, we take a step towards overcoming these difficulties. We present a method for global-likelihood optimization of a Markov random field language model exploiting long-range contexts in time independent of the corpus size. We take a variational approach to optimizing the likelihood and exploit underlying symmetries to greatly simplify learning. We demonstrate the efficiency of this method both for language modelling and for part-of-speech tagging.} }
Endnote
%0 Conference Paper %T A Fast Variational Approach for Learning Markov Random Field Language Models %A Yacine Jernite %A Alexander Rush %A David Sontag %B Proceedings of the 32nd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2015 %E Francis Bach %E David Blei %F pmlr-v37-jernite15 %I PMLR %P 2209--2217 %U https://proceedings.mlr.press/v37/jernite15.html %V 37 %X Language modelling is a fundamental building block of natural language processing. However, in practice the size of the vocabulary limits the distributions applicable for this task: specifically, one has to either resort to local optimization methods, such as those used in neural language models, or work with heavily constrained distributions. In this work, we take a step towards overcoming these difficulties. We present a method for global-likelihood optimization of a Markov random field language model exploiting long-range contexts in time independent of the corpus size. We take a variational approach to optimizing the likelihood and exploit underlying symmetries to greatly simplify learning. We demonstrate the efficiency of this method both for language modelling and for part-of-speech tagging.
RIS
TY - CPAPER TI - A Fast Variational Approach for Learning Markov Random Field Language Models AU - Yacine Jernite AU - Alexander Rush AU - David Sontag BT - Proceedings of the 32nd International Conference on Machine Learning DA - 2015/06/01 ED - Francis Bach ED - David Blei ID - pmlr-v37-jernite15 PB - PMLR DP - Proceedings of Machine Learning Research VL - 37 SP - 2209 EP - 2217 L1 - http://proceedings.mlr.press/v37/jernite15.pdf UR - https://proceedings.mlr.press/v37/jernite15.html AB - Language modelling is a fundamental building block of natural language processing. However, in practice the size of the vocabulary limits the distributions applicable for this task: specifically, one has to either resort to local optimization methods, such as those used in neural language models, or work with heavily constrained distributions. In this work, we take a step towards overcoming these difficulties. We present a method for global-likelihood optimization of a Markov random field language model exploiting long-range contexts in time independent of the corpus size. We take a variational approach to optimizing the likelihood and exploit underlying symmetries to greatly simplify learning. We demonstrate the efficiency of this method both for language modelling and for part-of-speech tagging. ER -
APA
Jernite, Y., Rush, A. & Sontag, D.. (2015). A Fast Variational Approach for Learning Markov Random Field Language Models. Proceedings of the 32nd International Conference on Machine Learning, in Proceedings of Machine Learning Research 37:2209-2217 Available from https://proceedings.mlr.press/v37/jernite15.html.

Related Material