Audio-to-Intent Using Acoustic-Textual Subword Representations from End-to-End ASR.
Pranay DighePrateeth NayakOggi RudovicErik MarchiXiaochuan NiuAhmed H. TewfikPublished in: ICASSP (2023)
Keyphrases
- end to end
- broadcast news
- automatic speech recognition
- acoustic features
- multimedia
- speech recognition
- multipath
- congestion control
- admission control
- audio signal
- high bandwidth
- wireless ad hoc networks
- speech signal
- content delivery
- hidden markov models
- internet protocol
- spoken document retrieval
- audio features
- ad hoc networks
- rate allocation
- real world
- out of vocabulary
- image sequences
- spontaneous speech
- packet loss rate