Foreign Language Analysis and Recognition (FLARE) Progress

Abstract

The following tasks were completed on Automatic Speech Recognition (ASR). Korean language models (LMs) were developed to reduce the number of Out-of-Vocabulary (OOV) words encountered by the recognizer. Levantine Arabic and Farsi ASR systems were trained on conversational telephone speech. Three different methods were investigated for combining Pashto ASR systems. Software was developed for training and evaluating hybrid deep neural network (DNN) hidden Markov model (HMM) speech recognition systems. An English ASR system was developed for the International Workshop on Spoken Language Translation (IWSLT) 2013 evaluation. Six different techniques were investigated for interpolating LM probabilities. Finally, English and Italian ASR systems were developed for the IWSLT 2014 evaluation.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Feb 01, 2015
Accession Number
ADA616191

Entities

People

  • Brian M. Ore
  • David M. Hoeferlin
  • Eric G. Hansen
  • Raymond E. Slyh
  • Stephen A. Thorn

Organizations

  • SRA International Inc

Tags

Communities of Interest

  • C4I
  • Energy and Power Technologies

DTIC Thesaurus Topics

  • Air Force
  • Air Force Research Laboratories
  • Automated Speech Recognition
  • Computer Languages
  • Computer Programming
  • Computer Science
  • Computers
  • Foreign Languages
  • Hidden Markov Models
  • Language
  • Linguistics
  • Machine Translation
  • Markov Models
  • Neural Networks
  • Operating Systems
  • Recognition
  • Recurrent Neural Networks

Readers

  • Computational Linguistics
  • Speech Processing/Speech Recognition.

Technology Areas

  • AI & ML
  • AI & ML - Machine Translation