Always-On Speech Recognition Using TrueNorth, a Reconfigurable, Neurosynaptic Processor.
Wei-Yu TsaiDavis R. BarchAndrew S. CassidyMichael V. DeBoleAlexander AndreopoulosBryan L. JacksonMyron D. FlicknerJohn V. ArthurDharmendra S. ModhaJohn SampsonVijaykrishnan NarayananPublished in: IEEE Trans. Computers (2017)
Keyphrases
- digital signal
- systolic array
- speech recognition
- parallel processing
- high speed
- reconfigurable architecture
- low cost
- single chip
- speech signal
- dialogue system
- functional units
- general purpose
- text to speech
- automatic speech recognition
- audio visual
- hardware implementation
- noisy environments
- endpoint detection
- spoken language
- single processor
- speaker recognition
- pattern recognition
- multiprocessor systems
- recognition engine
- computation intensive
- multi objective evolutionary
- functional verification
- hearing impaired
- power reduction
- instruction set
- broadcast news
- parallel processors
- computer architecture
- data flow