A chip architecture claimed to speed up speech recognition by
ten times over general purpose PC processors has been developed by
Carnegie-Mellon University (CMU).
“Whether running an enterprise-class voice call-in centre or
decoding individual words on a cell phone, all of today’s serious
speech recognisers exist as software running on some processor.
That’s terribly limiting,” said Professor Rob A. Rutenbar, who
leads the In Silico Vox project at CMU.
“Moving these computations directly into silicon means we can
perform recognition dramatically faster, cheaper, and better for
both commercial and homeland security tasks,” he said.
CMU’s architecture is exclusively dedicated to speech recognition
and performs it more efficiently than current PCs. The US Homeland
Security department is said to be very keen on CMU’s programme
because it will help them ‘audio-mine’, and translate huge streams
of incoming audio data, sifting for critical words or phrases.