Quoth Laslo Hunhold: > You'll not get around having to rely on a pre-trained neuronal network > unless you manage to formalize speech (partially done) or intelligence > (currently not done and probably impossible). > In this regard, personal assistants will by definition be bound to > centralized services, and if you ever think about solving it in open > source, it will have to be distributed like Diaspora or something > comparable. If you design it as a "single instance" from the get-go, it > will fail in the long run.
Well, training such a neural network would certainly require a lot of computer power and resources, but unless I'm mistaken I don't think running a query through it should be challenging for a non- internet connected computer. I haven't used speech recognition neural network stuff before, but I'd be surprised if it needed more space and compute power than were available on a smartphone. I don't really see the point of these assistants, but I certainly wouldn't consider one with the current architecture of "send every sound to large SF corporations."
