# US 2004/0083098 A1 · Method of performing speech recognition across a network

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2004/0083098 A1
- Document ID: US20040083098A1
- Status: Published application
- Kind: A1
- Application number: 10/686,927
- Filed: 2003-10-15
- Published: 2004-04-29
- Earliest claimed: 1997-03-24
- Assignees: Sensory, Incorporated
- Inventors: Todd F. Mozer; Forrest S. Mozer
- Web page: https://patentgrep.com/patent/US20040083098A1

## CPC

- G04G 21/06: using voice
- G10L 15/285: {Memory allocation or algorithm optimisation to reduce hardware requirements}
- G04G: ELECTRONIC TIME-PIECES
- G10L: SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING

## Abstract

Embodiments of the present invention include a method of performing speech recognition across a network. In one embodiment, the method includes providing, from a server to a first computer, sets of data to recognize spoken utterances from corresponding limited sets of candidate utterances, and supplying different sets of said data from the server to the first computer to recognize different spoken utterances from corresponding limited sets of candidate utterances at different times in response to different user interactions.

## Claims

### Claim 1

A method of performing speech recognition across a network comprising:
downloading first recognition information from a remote server to a first computer to recognize a first plurality of words;
programming the first computer with the first recognition information to recognize the first plurality of words;
receiving at least one of the first plurality of words in the first computer;
generating first recognition results in response to receiving said at least one of the first plurality of words;
downloading second recognition information from the remote server to the first computer to recognize a second plurality of words, wherein the second recognition information is selected based on the first recognition results; and
programming the first computer with the second recognition information to recognize a second plurality of words.

### Claim 2 (depends on 1)

The method of claim 1 wherein the first computer is connected to the server over an internet.

### Claim 3 (depends on 2)

The method of claim 2 wherein the first and second recognition information is downloaded from a internet web site.

### Claim 4 (depends on 1)

The method of claim 1 wherein the first computer is connected to the server over an intranet.

### Claim 5 (depends on 4)

The method of claim 4 wherein the first and second recognition information is downloaded from a intranet web site.

### Claim 6 (depends on 1)

The method of claim 1 wherein the first computer is connected to the server over a local network.

### Claim 7 (depends on 1)

The method of claim 1 wherein the first computer includes a software recognition engine.

### Claim 8 (depends on 7)

The method of claim 7 wherein the software recognition engine runs in a general purpose microprocessor.

### Claim 9 (depends on 1)

The method of claim 1 wherein recognition is performed using speaker-independent speech recognition.

### Claim 10 (depends on 9)

The method of claim 9 wherein the first and second recognition information comprise neural network weights.

### Claim 11

A method of performing speech recognition across a network comprising:
providing, from a server to a first computer, sets of data to recognize spoken utterances from corresponding limited sets of candidate utterances; and
supplying different sets of said data from the server to the first computer to recognize different spoken utterances from corresponding limited sets of candidate utterances at different times in response to different user interactions.

### Claim 12 (depends on 11)

The method of claim 11 wherein the first computer is connected to the server over an internet.

### Claim 13 (depends on 12)

The method of claim 12 wherein the first and second information is downloaded from a internet web site.

### Claim 14 (depends on 11)

The method of claim 11 wherein the first computer is connected to the server over an intranet.

### Claim 15 (depends on 14)

The method of claim 14 wherein the first and second information is downloaded from a intranet web site.

### Claim 16 (depends on 11)

The method of claim 11 wherein the first computer is connected to the server over a local network.

### Claim 17 (depends on 11)

The method of claim 11 wherein the first computer includes a software recognition engine.

### Claim 18 (depends on 17)

The method of claim 17 wherein the software recognition engine runs in a general purpose microprocessor.

### Claim 19 (depends on 11)

The method of claim 11 wherein recognition is performed using speaker-independent speech recognition.

### Claim 20 (depends on 19)

The method of claim 19 wherein the first and second information comprise neural network weights.

### Claim 21 (depends on 11)

The method of claim 11 further comprising prompting a user to input a first spoken utterance corresponding to a first limited set of candidate utterances and prompting a user to input a second spoken utterance corresponding to a second limited set of candidate utterances.

### Claim 22 (depends on 11)

The method of claim 11 wherein said data further includes synthesis data.

### Claim 23 (depends on 11)

The method of claim 11 wherein said data further includes video data.

### Claim 24 (depends on 11)

The method of claim 11 wherein recognition is performed using speaker-independent speech recognition.

### Claim 25 (depends on 24)

The method of claim 24 wherein the first and second information comprise neural network weights.

### Claim 26

A system for performing speech recognition across a network comprising:
a server including information to recognize a plurality of spoken utterances; and
a first computer including a recognition engine, wherein the first computer is coupled to the server by said network,
wherein the server supplies different sets of information to the first computer to recognize different spoken utterances from corresponding limited sets of candidate utterances at different times in response to different user interactions.

### Claim 27 (depends on 26)

The system of claim 26 wherein the server supplies the first computer with first information for recognizing a spoken utterance from a first limited set of candidate utterances, and the first computer is programmed with the first information to recognize the first limited set of candidate utterances.

### Claim 28 (depends on 27)

The system of claim 27 wherein the first computer generates first recognition results in response to receiving a spoken utterance from the first limited set of candidate utterances.

### Claim 29 (depends on 28)

The system of claim 28 wherein the server supplies the first computer with second information for recognizing a spoken utterance from a second limited set of candidate utterances, wherein the second information is selected based on the first recognition results, and the first computer is programmed with the second information to recognize the second limited set of candidate utterances.

### Claim 30 (depends on 26)

The system of claim 26 wherein the first computer is connected to the server over an internet.

### Claim 31 (depends on 26)

The system of claim 26 wherein the first computer is connected to the server over an intranet.

### Claim 32 (depends on 26)

The system of claim 26 wherein the first computer is connected to the server over a local network.

### Claim 33 (depends on 26)

The system of claim 26 wherein the recognition engine comprises software running in a general purpose microprocessor.

### Claim 34 (depends on 26)

The system of claim 26 wherein recognition is performed using speaker-independent speech recognition.

### Claim 35 (depends on 34)

The system of claim 34 wherein the recognition information comprises neural network weights.
