Matches in SemOpenAlex for { <https://semopenalex.org/work/W2489246890> ?p ?o ?g. }
Showing items 1 to 64 of
64
with 100 items per page.
- W2489246890 abstract "We applied a combination of artificial intelligence (AI) techniques with digital signal processing and statistical methods to enhance speech recognition. We implemented a hybrid system which used a rule-based expert system to create Conceptual Dependency (CD) representations of the spoken input. Conceptual Dependencies are used for natural language understanding of written text, but until now have not been applied to speech recognition. Our hybrid system used a three-step process. First, we implemented continuous speech recognition using a keyword spotting system, based on Hidden Markov Models. Then, the recognized keywords were used to search for the CD representation. Finally, the gaps in the CD representation were used as contextual cues for reinterpreting the speech input and so increase the speech recognition accuracy.We also implemented a sentence speech recognition system based on Multi-Section Vector Quantization, CD representations, and context. The recognized sentences were used to create CD representations, which were then combined with contextual knowledge to reinterpret the speech input and increase the recognition accuracy. The context was represented using Scripts, which describe possible sequences of events that an actor or entity may perform under certain conditions. Using context representation reduced the speech recognition errors in our application by 31%.The application chosen was software agent navigation through a virtual environment via vocal commands. The software agent was represented by a Virtual Robot (VIRBOT) and the environment was the representation of a simple house. There were a number of objects in the house and the VIRBOT could be directed to them and could plan a series of actions to achieve a goal. The VIRBOT understood commands such as Robot, give the newspaper to the father, Robot, where is the tool box?, Robot, bring it, etc. Feedback was provided by synthesized speech announcing the movements of the VIRBOT; for example the robot says I found the newspaper, when it finds it. Another type of feedback was provided visually to the user, who could see the VIRBOT as it moved. This application may be used to develop and test simulated Robot behaviors, before construction and programming of a real Robot.This work used the best current methods for speech recognition, and integrated them with semantic representations (CD's) and contextual representations (scripts.) It has been shown that meaning and context can improve speech recognition accuracy considerably, in a real time environment." @default.
- W2489246890 created "2016-08-23" @default.
- W2489246890 creator A5007036872 @default.
- W2489246890 date "1996-10-03" @default.
- W2489246890 modified "2023-09-23" @default.
- W2489246890 title "A hybrid system with symbolic AI and statistical methods for speech recognition" @default.
- W2489246890 hasPublicationYear "1996" @default.
- W2489246890 type Work @default.
- W2489246890 sameAs 2489246890 @default.
- W2489246890 citedByCount "8" @default.
- W2489246890 countsByYear W24892468902013 @default.
- W2489246890 crossrefType "journal-article" @default.
- W2489246890 hasAuthorship W2489246890A5007036872 @default.
- W2489246890 hasConcept C154945302 @default.
- W2489246890 hasConcept C155635449 @default.
- W2489246890 hasConcept C195324797 @default.
- W2489246890 hasConcept C199360897 @default.
- W2489246890 hasConcept C204321447 @default.
- W2489246890 hasConcept C23224414 @default.
- W2489246890 hasConcept C2777530160 @default.
- W2489246890 hasConcept C28490314 @default.
- W2489246890 hasConcept C41008148 @default.
- W2489246890 hasConcept C54953205 @default.
- W2489246890 hasConcept C61328038 @default.
- W2489246890 hasConcept C61423126 @default.
- W2489246890 hasConceptScore W2489246890C154945302 @default.
- W2489246890 hasConceptScore W2489246890C155635449 @default.
- W2489246890 hasConceptScore W2489246890C195324797 @default.
- W2489246890 hasConceptScore W2489246890C199360897 @default.
- W2489246890 hasConceptScore W2489246890C204321447 @default.
- W2489246890 hasConceptScore W2489246890C23224414 @default.
- W2489246890 hasConceptScore W2489246890C2777530160 @default.
- W2489246890 hasConceptScore W2489246890C28490314 @default.
- W2489246890 hasConceptScore W2489246890C41008148 @default.
- W2489246890 hasConceptScore W2489246890C54953205 @default.
- W2489246890 hasConceptScore W2489246890C61328038 @default.
- W2489246890 hasConceptScore W2489246890C61423126 @default.
- W2489246890 hasLocation W24892468901 @default.
- W2489246890 hasOpenAccess W2489246890 @default.
- W2489246890 hasPrimaryLocation W24892468901 @default.
- W2489246890 hasRelatedWork W1560013842 @default.
- W2489246890 hasRelatedWork W193854645 @default.
- W2489246890 hasRelatedWork W1963720326 @default.
- W2489246890 hasRelatedWork W1969038863 @default.
- W2489246890 hasRelatedWork W198511600 @default.
- W2489246890 hasRelatedWork W2135399268 @default.
- W2489246890 hasRelatedWork W2148876114 @default.
- W2489246890 hasRelatedWork W2152837596 @default.
- W2489246890 hasRelatedWork W2294568999 @default.
- W2489246890 hasRelatedWork W2548283226 @default.
- W2489246890 hasRelatedWork W2963288440 @default.
- W2489246890 hasRelatedWork W3016367418 @default.
- W2489246890 hasRelatedWork W3164394706 @default.
- W2489246890 hasRelatedWork W758479746 @default.
- W2489246890 hasRelatedWork W1912342180 @default.
- W2489246890 hasRelatedWork W2091056747 @default.
- W2489246890 hasRelatedWork W2108421539 @default.
- W2489246890 hasRelatedWork W2269090647 @default.
- W2489246890 hasRelatedWork W2508928779 @default.
- W2489246890 hasRelatedWork W2513977593 @default.
- W2489246890 isParatext "false" @default.
- W2489246890 isRetracted "false" @default.
- W2489246890 magId "2489246890" @default.
- W2489246890 workType "article" @default.