Home › Categories › Technology› Segment-Based Acoustic Models With Multi-level Search Algorithms For Continuous Speech Recognition

Segment-Based Acoustic Models With Multi-level Search Algorithms For Continuous Speech Recognition.pdf

H92-1100.pdf
Preview of Segment-Based Acoustic Models with Multi-level Search Algorithms for Continuous Speech Recognition
🔗 Source: aclanthology.org
📊 Size: 92 KB
👤 Author: Marl Ostendorf; J. Robin Rohlicek
⬇️ Downloads: 51

Summary

Developed robust context modeling for stochastic segment models using tied covariance distributions, reducing error rate by a factor of two on the RM Oct 89 test set. Improved the BBN-ItMM/BU-SSM combined system to 3.3% word error. Determined that linear models have predictive power similar to non-linear models of cepstra within segments. Evaluated the dynamical system model in phoneme recognition, outperforming the independent-frame model on the TIMIT corpus. Reformulated the recognition problem as a classification and segmentation scoring problem, allowing more general types of classifiers and non-traditional feature analysis. Demonstrated that for equivalent feature sets and context-independent models, the two methods give similar results. Investigated duration models conditioned on speaking rate and pre-pause location, improving performance by increasing the weight of duration. Analyzed the behavior of recognition error over the weight space for HMM and SSM scores in the N-best rescoring paradigm, addressing the problem of local optima with a grid-based search. Discovered a significant mismatch problem between training and test data. Extended Bayesian techniques for speaker adaptation, achieving a 16% reduction in error using 3 minutes of speech with simple mean adaptation techniques. Developed a multi-level stochastic model of speech that can take advantage of multi-rate signal analysis, evaluating the model for the two-level case with cepstral features and showing improved performance over a single-level model.

Download File

Description

Ostendorf and Rohlicek (Boston U, BBN) aim to enhance speaker-independent continuous speech recognition via stochastic, segment-based acoustic models and efficient, multi-level search algorithms, funded by DARPA and NSF.

Technical Information

  • File Format: PDF
  • File Size: 92 KB
  • Pages: 1
  • Language: EN
  • Author: Marl Ostendorf; J. Robin Rohlicek
  • Total Downloads: 51
  • Last Updated: 2 hours ago

Document Overview

This PDF document about Segment-Based Acoustic Models with Multi-level Search Algorithms for Continuous Speech Recognition provides comprehensive information and guidance. Whether you're a beginner or advanced user, this resource offers valuable insights into Segment-Based Acoustic Models with Multi-level Search Algorithms for Continuous Speech Recognition.

Related Topics

If you're interested in Segment-Based Acoustic Models with Multi-level Search Algorithms for Continuous Speech Recognition, you might also want to explore:

Segment-Based guideSegment-Based tutorialAcoustic guideAcoustic tutorialModels guideModels tutorialwith guidewith tutorialMulti-level guideMulti-level tutorialSearch guideSearch tutorialAlgorithms guideAlgorithms tutorialContinuous guideContinuous tutorialSpeech guideSpeech tutorialRecognition guideRecognition tutorial

Related Categories

Business & Finance
41204 documents
Education & Learning
22711 documents
Health & Wellness
8956 documents

Related Documents

Preview of pdf
PDF file

pdf

85 KBVassilios Digalakis; Hy Murveit; Mitch WeintraubEN1 page
The project aims to develop acoustic modeling techniques for speech recognition, focusing on relaxing hidden Markov model's independence assumptions.
Download
Preview of Speech Recognition Models
PDF file

Speech Recognition Models

2.28 MBEN19 pages
Researchers propose attention-based models for speech recognition, extending existing attention mechanisms to handle long and noisy input sequences,...
Download
Preview of Evaluation of 2D Still-Image Face Recognition Algorithms
PDF file

Evaluation of 2D Still-Image Face Recognition Algorithms

2.69 MBpatrickEN61 pages
It evaluates performance across diverse populations and operational scenarios, aiming to advance facial recognition technology's accuracy and...
Download
Preview of <think>
PDF file

<think>

444 KBRajmund Nagy, Taras Kucherenko, Birger Moell, André Pereira, Hedvig Kjellström, and Ulysses BernardeEN3 pages
Let me start by reading through the provided information. The title is "A Framework for Integrating Gesture Generation Models into Interactive...
Download
Preview of Heuristic Search Algorithm
PDF file

Heuristic Search Algorithm

197 KBBradford Larsen, Ethan Burns, Wheeler Ruml, Robert C. HolteEN7 pages
The authors propose a novel hierarchical heuristic search algorithm called Switchback that efficiently uses abstraction to derive admissible...
Download

Download Segment-Based Acoustic Models with Multi-level Search Algorithms for Continuous Speech Recognition eBooks for free and learn more about Segment-Based Acoustic Models with Multi-level Search Algorithms for Continuous Speech Recognition. These books contain exercises and tutorials to improve your practical skills, at all levels!

Not satisfied with this document? We have related documents to Segment-Based Acoustic Models with Multi-level Search Algorithms for Continuous Speech Recognition, try searching with similar keywords: Segment-Based Acoustic Models with Multi-level Search Algorithms for Continuous Speech Recognition, "Hinder och möjliggörare för 1.5°-livsstilar: Ytliga och djupgående strukturella faktorer som påverkar potentialen för hållbar k, Ändring av genomföranderam för en europeisk plattform för utbyte av balansenergi från frekvensåterställn ingsreserver med manuell, Rekommendationer för vaccination mot covid-19 för särskilda grupper av barn -, förstudie för att utvärdera förutsättningarna att genom en innovationsupphandli ng utveckla en drifttjänst för geoenergilager, Självkänsla och KBT ‐ Påverkas självkänslan vid KBT för depression och ångesttillstånd?Se lf‐esteem and CBT ‐ How does CBT for de, Matglädje för alla: en guide till rätt konsistens för olika behov, decodeur 7 segment listes des fichiers pdf decodeur 7 segment

You can download PDF versions of the user's guide, manuals and ebooks about Segment-Based Acoustic Models with Multi-level Search Algorithms for Continuous Speech Recognition, you can also find and download for free A free online manual (notices) with beginner and intermediate, Downloads Documentation, You can download PDF files (or DOC and PPT) about Segment-Based Acoustic Models with Multi-level Search Algorithms for Continuous Speech Recognition for free, but please respect copyrighted ebooks.