Unlock instant, AI-driven research and patent intelligence for your innovation.

Transcription correction using multi-token structures

A marking and marking technology, applied in speech analysis, speech recognition, instruments, etc., can solve problems such as imperfect speech recognition

Active Publication Date: 2020-08-07
MICROSOFT TECH LICENSING LLC
View PDF6 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Speech recognition isn't perfect, and every user understands that occasional recognition errors are a reality

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Transcription correction using multi-token structures
  • Transcription correction using multi-token structures
  • Transcription correction using multi-token structures

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0014] A method of correcting speech recognition errors may use a word confusion network that may provide alternatives for certain words once the user indicates that a hypothesis (eg, result) provided to the user is not what the user wants. However, in general, word confusion networks (WCNs) do not address the problem of alternatives or corrections across multiple words or nodes of a WCN. An additional challenge comes from the fact that speech recognition occurs at the lexical level, and thus WCNs are generated at the lexical level, where the text presented to the user contains tokens as a result of text normalization on the lexical output. Thus, common WCNs may struggle to handle corrections in the presence of altered words associated with spoken utterances.

[0015] Examples of the present disclosure describe the generation of multi-arc token-level confusion networks that represent hypotheses for recognition results of spoken utterances to improve the ability to return to us...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

Examples of the present disclosure describe generation of a multi-arc confusion network to improve, for example, an ability to return alternatives to output generated. A confusion network comprising token representations of lexicalized hypotheses and normalized hypotheses is generated. Each arc of the confusion network represents a token of a lexicalized hypothesis or a normalized hypothesis. Theconfusion network is transformed into a multi-arc confusion network, wherein the transforming comprising realigning at least one token of the confusion network to span multiple arcs of the confusion network. Other examples are also described.

Description

[0001] Description of divisional application [0002] This application is a divisional application of a Chinese invention patent application with an application date of January 22, 2016, an application number of 201680005243.1, and an invention title of "Transcription correction using a multi-label structure". Background technique [0003] Advances in automatic speech recognition (ASR) have led to increased interest in spoken language understanding (SLU). A challenge in large-vocabulary spoken language understanding is the robustness to compensate for ASR errors. Speech recognition isn't perfect, and every user understands that occasional recognition errors are a reality. From a user's perspective, the ease of correction of recognition errors has a significant impact on the user's overall experience when a speech recognition application or program is used. This is about this general technical environment for which this application is aimed. Contents of the invention [0...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Applications(China)
IPC IPC(8): G10L15/30G10L15/22G10L25/33G10L15/187G10L15/08G10L15/02G10L15/01G06F40/40G06F40/284G10L15/197
CPCG10L15/083G10L15/187G10L15/22G10L2015/221G10L15/197G06F40/284G10L15/30G10L25/33G10L15/02G06F40/40G10L15/01
Inventor M·莱维特U·奥泽特姆S·帕撒萨拉塞P·瓦拉德哈拉简K·拉古纳森I·阿方索
Owner MICROSOFT TECH LICENSING LLC