Skip this Video
Utilizing Commonsense Reasoning to Improve Voice Recognition
Utilizing Commonsense Reasoning to Improve Voice Recognition - PowerPoint PPT Presentation
Discourse Recognition. Disambiguate phonetically comparable words Improve blunder rectification interfaces Improve general acknowledgment exactness. By incorporating the Microsoft Speech API and ConceptNet, we want to:. Discourse Recognition. Here is the means by which our model is going to work:. I was at the school a day or two ago. I strolled up to the(text being managed shows up here).
Uploaded on |
456 Views
Presentation Transcript
Speech Recognition By integrating the Microsoft Speech API and ConceptNet, we hope to: • Disambiguate phonetically similar words • Improve errorcorrection interfaces • Improve overall recognition accuracy
Speech Recognition Here is how our prototype is going to work: I was at the school the other day. I walked up to the (text being dictated appears here)
principal Speech Recognition User Speaks: I was at the school the other day. I walked up to the
Voice recognition engine suggestions Speech Recognition Speech is Processed: I was at the school the other day. I walked up to the principle principal the principle of of the principal
Words in ConceptNet for “school” Speech Recognition The Context is analyzed using ConceptNet I was at the school the other day. I walked up to the principle principal the principle of of the principal chalkboard textbook teacher principal
The voice recognition input is intelligently sorted by how much each word makes sense, contextually Speech Recognition Options are calculated and ranked: I was at the school the other day. I walked up to the principal principal principle the principle of of the principal principle principal the principle of of the principal chalkboard textbook teacher principal