Conditional random fields for integrating local discriminative classifiers

DMCA

Conditional random fields for integrating local discriminative classifiers (2008)

Cached

by Jeremy Morris , Student Member , Eric Fosler-lussier , Senior Member

Venue:	Audio, Speech, and Language Processing, IEEE Transactions on
Citations:	18 - 2 self

BibTeX

@ARTICLE{Morris08conditionalrandom,
    author = {Jeremy Morris and Student Member and Eric Fosler-lussier and Senior Member},
    title = {Conditional random fields for integrating local discriminative classifiers},
    journal = {Audio, Speech, and Language Processing, IEEE Transactions on},
    year = {2008},
    pages = {617--628}
}

OpenURL

Abstract

Abstract—Conditional random fields (CRFs) are a statistical framework that has recently gained in popularity in both the automatic speech recognition (ASR) and natural language processing communities because of the different nature of assumptions that are made in predicting sequences of labels compared to the more traditional hidden Markov model (HMM). In the ASR community, CRFs have been employed in a method similar to that of HMMs, using the sufficient statistics of input data to compute the probability of label sequences given acoustic input. In this paper, we explore the application of CRFs to combine local posterior estimates provided by multilayer perceptrons (MLPs) corresponding to the frame-level prediction of phone classes and phonological attribute classes. We compare phonetic recognition using CRFs to an HMM system trained on the same input features and show that the monophone label CRF is able to achieve superior performance to a monophone-based HMM and performance comparable to a 16 Gaussian mixture triphone-based HMM; in both of these cases, the CRF obtains these results with far fewer free parameters. The CRF is also able to better combine these posterior estimators, achieving a substantial increase in performance over an HMM-based triphone system by mixing the two highly correlated sets of phone class and phonetic attribute class posteriors. Index Terms—Automatic speech recognition (ASR), random fields. I.

Keyphrases

local discriminative classifier conditional random field phone class different nature input data free parameter automatic speech recognition hmm-based triphone system abstract conditional random field phonetic recognition multilayer perceptrons label sequence random field correlated set superior performance monophone-based hmm gaussian mixture triphone-based hmm input feature substantial increase monophone label crf natural language processing community asr community frame-level prediction index term automatic speech recognition phonetic attribute class posterior local posterior estimate sufficient statistic posterior estimator hmm system phonological attribute class traditional hidden markov model acoustic input statistical framework

DMCA