|
Authored by: Anonymous on Friday, July 20 2012 @ 04:18 PM EDT |
It depends on the level of skill of one of ordinary skill in the art and upon
where the novelty lies.
If one of ordinary skill would understand how to implement each of those steps,
and if those steps have never been combined in that way before then it is
probably enough.
Since you were able to rattle them off like that, I suspect they are not new and
that the novelty would have to lie in the details of a particular implementation
of at least one of those steps. In that case, more would be required and the
claims would have to be more specific.
Also, Hal did more than just parse phonemes... [ Reply to This | Parent | # ]
|
|
Authored by: Anonymous on Sunday, July 22 2012 @ 12:01 PM EDT |
To me, you provided a sketch of an algorithm for voice recognition. The
function-interface you describe takes recorded audio as input and has as output
either a transcription of the utterance or a message that should be presented to
the user, followed by processing the next utterance.
But you went further, and (barely) started to describe the algorithm: raw audio
-> phonetic parsing -> assemble words -> disambiguate using semantics
-> prompt if too ambiguous. But accomplishing these steps in speech
recognition are highly nontrivial. Someone skilled in the art of speech rec
would know at least the canonical implementation involving creating acoustic and
language models (for phonetic and semantic parsing respectively) and would
create a decoder that uses those data structures.
I'd hope that any patent in this area would show specifically where the
innovation is, what new data structures and sequences of steps on those data
structures were invented. In other words, present details of the secret sauce.
Then it would have to be non-obvious to practicioners and original.
Also note that providing this _appropriate_ level of detail for a patent (data
structures and the sequence of operations on the data structures) looks
suspiciously like math.
[ Reply to This | Parent | # ]
|
|
|
|
|