As a long time advocate of voice recognition I was quite pleased to see Google receive a patent for a voice recognition front end to web searching. The implications are huge, not the least of which would be voice searching on mobile phones or other mobile devices. Speech recognition has made great strides in recent years as any Tablet PC owner can attest but those improvements are largely due to the ability to interpret words in context. The speech engines try to determine what word was spoken by examining the words and phrases around the particular word which increases the accuracy quite a bit. This context interpretation is something that is not that useful with search terms as they are usually too short to provide any clues. Google’s patent indicates they have gotten around that obstacle which is very exciting for us speech buffs. Google has given no indication how or when this might be implemented but let’s hope it will show up soon.
(via ars technica)