# US 2021/0081609 A1 · Streamlining dialog processing using integrated shared resources

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

- Number: US 2021/0081609 A1
- Document ID: US20210081609A1
- Status: Published application
- Kind: A1
- Application number: 16/927,357
- Filed: 2020-07-13
- Published: 2021-03-18
- Earliest claimed: 2019-09-12
- Assignees: Oracle International Corporation
- Applicants: Oracle International Corporation
- Inventors: Mark Edward Johnson
- Web page: https://patentgrep.com/patent/US20210081609A1

## CPC

- G06F 40/242: Dictionaries
- G06F 40/30: Semantic analysis
- G10L 15/1822: {Parsing for meaning understanding}
- G10L 15/187: Phonemic context, e.g. pronunciation rules, phonotactical constraints or phoneme n-grams
- G10L 15/22: Procedures used during a speech recognition process, e.g. man-machine dialogue
- G10L 15/28: Constructional details of speech recognition systems
- G06F 40/295: Named entity recognition
- G10L 2015/226: {using non-speech characteristics}
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G10L: SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING

## Abstract

Techniques for reducing memory and processing resources used by a dialog system by sharing resources between pipelined processes of the dialog system. An integrated shared dictionary is constructed for concurrent use by automated speech recognition (ASR) and natural language understanding (NLU) subsystems of the dialog system. The integrated shared dictionary comprises multiple entries, with each entry comprising first information that is used by the ASR subsystem, second information used by the NLU subsystem, and information correlating the first information and the second information. The ASR subsystem uses the integrated shared dictionary to identify a dictionary entry containing a set of words corresponding to speech input. The dictionary entry information is communicated to the NLU subsystem, which uses the entry to generate a meaning representation for the speech input. Sharing the dictionary between the subsystems results in significant savings in memory resources used by the dialog system and speeds processing.

## Claims

### Claim 1

In a dialog system implemented using one or more processors, a computer-implemented method comprising:
receiving, by a first subsystem of the dialog system, speech input;
processing, by the first subsystem, the speech input to identify a pronunciation corresponding to the speech input;
searching, by the first subsystem, a dictionary stored in a memory to find a matching entry in the dictionary corresponding to the identified pronunciation, the dictionary comprising a plurality of entries, each entry in the plurality of entries storing a set of one or more words mapped to a pronunciation stored in the entry, each entry in the plurality of entries further comprising additional information usable for determining a meaning representation for the set of one or more words stored by the entry;
communicating, by the first subsystem to a second subsystem of the dialog system, information identifying the matching entry; and
generating, by the second subsystem, a meaning representation for the speech input using the information identifying the entry, the meaning representation identifying an intent associated with the speech input, wherein the generating comprises using, by the second subsystem, the additional information stored in the matching entry in the dictionary.

### Claim 2 (depends on 1)

The method of claim 1, wherein:
the first subsystem is an Automated Speech Recognition (ASR) subsystem; and
the second subsystem is a Natural Language Understanding (NLU) subsystem.

### Claim 3 (depends on 1)

The method of claim 1, wherein:
the information indicating the entry comprises a pointer to the additional information; and
the second subsystem of the dialog system uses the pointer to retrieve the additional information.

### Claim 4 (depends on 1)

The method of claim 1, wherein:
the additional information comprises a named entity type corresponding to the set of words; and
generating the meaning representation comprises linking the named entity type to the set of words.

### Claim 5 (depends on 1)

The method of claim 1, wherein:
the set of words is a first set of words; and
the method further comprises:
identifying, by the first subsystem of the dialog system using the dictionary, a second set of words;
outputting, by the first subsystem of the dialog system to the second subsystem of the dialog system, the second set of words; and
determining, by the second subsystem of the dialog system, a context by associating the words, of the first set of words and the second set of words, with one another using a machine learning model.

### Claim 6 (depends on 1)

The method of claim 1, further comprising:
outputting, by the second subsystem of the dialog system to a third subsystem of the dialog system, the generated meaning representation.

### Claim 7 (depends on 1)

The method of claim 1, further comprising:
identifying information used by the first subsystem;
identifying information used by the second subsystem;
determining information correlating the information used by the first subsystem and the information used by the second subsystem;
generating the dictionary used by the first subsystem, the information used by the second subsystem, and the correlating information; and
making the dictionary accessible by the first subsystem and the second subsystem.

### Claim 8

A non-transitory computer-readable memory storing a plurality of instructions executable by one or more processors, the plurality of instructions comprising instructions that when executed by the one or more processors cause the one or more processors to perform processing comprising:
receiving, by a first subsystem of a dialog system, speech input;
processing, by the first subsystem, the speech input to identify a pronunciation corresponding to the speech input;
searching, by the first subsystem, a dictionary stored in a memory to find a matching entry in the dictionary corresponding to the identified pronunciation, the dictionary comprising a plurality of entries, each entry in the plurality of entries storing a set of one or more words mapped to a pronunciation stored in the entry, each entry in the plurality of entries further comprising additional information usable for determining a meaning representation for the set of one or more words stored by the entry;
communicating, by the first subsystem to a second subsystem of the dialog system, information identifying the matching entry; and
generating, by the second subsystem, a meaning representation for the speech input using the information identifying the entry, the meaning representation identifying an intent associated with the speech input, wherein the generating comprises using, by the second subsystem, the additional information stored in the matching entry in the dictionary.

### Claim 9 (depends on 8)

The non-transitory computer-readable memory of claim 8, wherein:
the first subsystem is an Automated Speech Recognition (ASR) subsystem; and
the second subsystem is a Natural Language Understanding (NLU) subsystem.

### Claim 10 (depends on 8)

The non-transitory computer-readable memory of claim 8, wherein:
the information indicating the entry comprises a pointer to the additional information; and
the second subsystem of the dialog system uses the pointer to retrieve the additional information.

### Claim 11 (depends on 8)

The non-transitory computer-readable memory of claim 8, wherein:
the additional information comprises a named entity type corresponding to the set of words; and
generating the meaning representation comprises linking the named entity type to the set of words.

### Claim 12 (depends on 8)

The non-transitory computer-readable memory of claim 8, wherein:
the set of words is a first set of words; and
the processing further comprises:
identifying, by the first subsystem of the dialog system using the dictionary, a second set of words;
outputting, by the first subsystem of the dialog system to the second subsystem of the dialog system, the second set of words; and
determining, by the second subsystem of the dialog system, a context by associating the words, of the first set of words and the second set of words, with one another using a machine learning model.

### Claim 13 (depends on 8)

The non-transitory computer-readable memory of claim 8, the processing further comprising:
outputting, by the second subsystem of the dialog system to a third subsystem of the dialog system, the generated meaning representation.

### Claim 14 (depends on 8)

The non-transitory computer-readable memory of claim 8, the processing further comprising:
identifying information used by the first subsystem;
identifying information used by the second subsystem;
determining information correlating the information used by the first subsystem and the information used by the second subsystem;
generating the dictionary used by the first subsystem, the information used by the second subsystem, and the correlating information; and
making the dictionary accessible by the first subsystem and the second subsystem.

### Claim 15

A dialog system comprising:
one or more processors;
a memory coupled to the one or more processors, the memory storing a plurality of instructions executable by the one or more processors, the plurality of instructions comprising instructions that when executed by the one or more processors cause the one or more processors to perform processing comprising:
receiving, by a first subsystem of the dialog system, speech input;
processing, by the first subsystem, the speech input to identify a pronunciation corresponding to the speech input;
searching, by the first subsystem, a dictionary stored in the memory to find a matching entry in the dictionary corresponding to the identified pronunciation, the dictionary comprising a plurality of entries, each entry in the plurality of entries storing a set of one or more words mapped to a pronunciation stored in the entry, each entry in the plurality of entries further comprising additional information usable for determining a meaning representation for the set of one or more words stored by the entry;
communicating, by the first subsystem to a second subsystem of the dialog system, information identifying the matching entry; and
generating, by the second subsystem, a meaning representation for the speech input using the information identifying the entry, the meaning representation identifying an intent associated with the speech input, wherein the generating comprises using, by the second subsystem, the additional information stored in the matching entry in the dictionary.

### Claim 16 (depends on 15)

The dialog system of claim 15, wherein:
the first subsystem is an Automated Speech Recognition (ASR) subsystem; and
the second subsystem is a Natural Language Understanding (NLU) subsystem.

### Claim 17 (depends on 15)

The dialog system of claim 15, wherein:
the information indicating the entry comprises a pointer to the additional information; and
the second subsystem of the dialog system uses the pointer to retrieve the additional information.

### Claim 18 (depends on 15)

The dialog system of claim 15, wherein:
the additional information comprises a named entity type corresponding to the set of words; and
generating the meaning representation comprises linking the named entity type to the set of words.

### Claim 19 (depends on 15)

The dialog system of claim 15, the processing further comprising:
outputting, by the second subsystem of the dialog system to a third subsystem of the dialog system, the generated meaning representation.

### Claim 20 (depends on 15)

The dialog system of claim 15, the processing further comprising:
identifying information used by the first subsystem;
identifying information used by the second subsystem;
determining information correlating the information used by the first subsystem and the information used by the second subsystem;
generating the dictionary used by the first subsystem, the information used by the second subsystem, and the correlating information; and
making the dictionary accessible by the first subsystem and the second subsystem.
