DISCRIMINATIVE SPOKEN TERM DETECTION WITH LIMITED DATA

Rohit Prabhavalkar, Joseph Keshet, Karen Livescu, Eric Fosler-Lussie

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

We study spoken term detection - the task of determining whether and where a given word or phrase appears in a given segment of speech - in the setting of limited training data. This setting is becoming increasingly important as interest grows in porting spoken term detection to multiple lowresource languages and acoustic environments. We propose a discriminative algorithm that aims at maximizing the area under the receiver operating characteristic curve, often used to evaluate the performance of spoken term detection systems. We implement the approach using a set of feature functions based on multilayer perceptron classifiers of phones and articulatory features, and experiment on data drawn from the Switchboard database of conversational telephone speech. Our approach outperforms a baseline HMM-based system by a large margin across a number of training set sizes.

Original languageEnglish
Title of host publication2012 Symposium on Machine Learning in Speech and Language Processing, MLSLP 2012
Pages22-25
Number of pages4
StatePublished - 2012
Externally publishedYes
Event2012 Symposium on Machine Learning in Speech and Language Processing, MLSLP 2012 - Portland, United States
Duration: 14 Sep 2012 → …

Conference

Conference2012 Symposium on Machine Learning in Speech and Language Processing, MLSLP 2012
Country/TerritoryUnited States
CityPortland
Period14/09/12 → …

Keywords

  • AUC
  • Spoken term detection
  • discriminative training
  • structural SVM

All Science Journal Classification (ASJC) codes

  • Artificial Intelligence
  • Human-Computer Interaction
  • Computer Vision and Pattern Recognition
  • Signal Processing
  • Linguistics and Language

Fingerprint

Dive into the research topics of 'DISCRIMINATIVE SPOKEN TERM DETECTION WITH LIMITED DATA'. Together they form a unique fingerprint.

Cite this