US 11,908,463 B1Granted

Multi-session context

AssigneeAmazon Technologies, Inc.

InventorsArjit Biswas, Shishir Bharathi, Anushree Venkatesh, Yun Lei, Ashish Kumar Agrawal, Siddhartha Reddy Jonnalagadda, Prakash Krishnan, Arindam Mandal, Raefer Christopher Gabriel, Abhay Kumar Jha, David Chi-Wai Tang, Savas Parastatidis

Life of this patent

Estimating the term…

patentgrep

Loading patent…

Abstract

Techniques for storing and using multi-session context are described. A system may store context data corresponding to a first interaction, where the context data may include action data, entity data and a profile identifier for a user. Later the stored context data may be retrieved during a second interaction corresponding to the entity of the second interaction. The second interaction may take place at a system different than the first interaction. The system may generate a response during the second interaction using the stored context data of the prior interaction.

Claims18

Dependent claims

A computer-implemented method comprising:

receiving, from a first device and at a first system, first audio data corresponding to a first spoken input, the first audio data associated with a profile identifier and a first session identifier;
determining, using the first audio data, first entity data corresponding to the first spoken input, the first entity data associated with the first session identifier;
receiving first context data associated with the profile identifier and relating to a first prior action at a second system different than the first system, the first context data including second entity data, the second entity data being related to the first entity data;
receiving second context data associated with the profile identifier and relating to a second prior action, a type of the second prior action being different than a type of the first prior action;
selecting the first context data, instead of the second context data, as corresponding to the first entity data, based on the first prior action and the second prior action;
processing, using a first dialog model, the first audio data, the first entity data and the first context data to determine a system response to the first spoken input;
sending, to the first device, output data corresponding to the system response;
receiving, from the first device, second audio data corresponding to a second spoken input responsive to the system response, the second audio data associated with the profile identifier and the first session identifier;
determining, using the second audio data, a third action to be performed using the first entity data and the second entity data;
and
storing third context data including the first entity data, the second entity data, the third action and the profile identifier.

8 more elements

A computer-implemented method comprising:

receiving, at a first system, first input data associated with a profile identifier;
determining first entity data corresponding to the first input data;
receiving first context data associated with the profile identifier and relating to a first prior action at a second system, the first context data including at least second entity data, the second entity data being related to the first entity data;
receiving second context data associated with the profile identifier and relating to a second prior action, a type of the second prior action being different than a type of the first prior action;
selecting the first context data, instead of the second context data, as corresponding to the first entity data, based on the first prior action and the second prior action;
processing, using a first dialog model, the first input data, the first entity data and the first context data;
and
determining, based on processing using the first dialog model, a first system response to the first input data.

5 more elements

A first system comprising:

at least one processor;
and
at least one memory including instructions that, when executed by the at least one processor, cause the first system to:
receive first input data associated with a profile identifier;
determine first entity data corresponding to the first input data;
receive first context data associated with the profile identifier and relating to a first prior action at a second system, the first context data including at least second entity data, the second entity data being related to the first entity data;
receive second context data associated with the profile identifier and relating to a second prior action, a type of the second prior action being different than a type of the first prior action;
selecting the first context data, instead of the second context data, as corresponding to the first entity data, based on the first prior action and the second prior action;
processing, using a first dialog model, the first input data, the first entity data and the first context data;
and
determine, based on processing using the first dialog model, a first system response to the first input data.

8 more elements

Dependencies are inferred from the claim text.

Facts

Application: 17/361,761Filed 29 Jun 2021
Granted: 20 Feb 2024
Term ends: Not estimatedThe source records don’t support an estimate.

Classification12 codes

G10L 15/183
Using natural language modelling: using context dependencies, e.g. language models
G06F 40/279
Natural language analysis: recognition of textual entities
G10L 15/1815
Using natural language modelling: semantic context, e.g. disambiguation of the recognition hypotheses based on word meaning
G10L 15/22
Speech recognition: procedures used during a speech recognition process, e.g. man-machine dialogue

Family

No related US applications are recorded in the indexed filing history.

Timeline

Citations

Open elsewhere

Building an agent or pulling many patents? Use the API or MCP server instead of scraping pages. Get an API key