US 12,597,421 B1Granted

Event timeline system

AssigneeAmazon Technologies, Inc.

InventorsShyam Sunder Kumar, Vandana Sathpathy, Aparna Kakarparti

Life of this patent

Estimating the term…

Loading patent…

Abstract

Devices and techniques are generally described for an event timeline system. In some examples, first prompt data including a first natural language input may be generated. An LLM may generate a first directive to send a first request to a first interface of a first computer-implemented system using the first prompt data. First result data may be received from the first computer-implemented system in response to the first request. The first result data may include a first time-stamped event associated with a first device. Second prompt data including the first natural language input and a representation of the first result data may be generated. The LLM may generate a first output action responsive to the first natural language input using the second prompt data.

Claims20

Dependent claims

A computer-implemented method comprising:

receiving a first natural language input;
generating first prompt data comprising the first natural language input and an account identifier;
generating, by a large language model (LLM) using the first prompt data, first action plan data, wherein the first action plan data comprises a directive to send a first application programming interface (API) request to a first API of a first computer-implemented system, wherein the first API request comprises the account identifier;
receiving first result data from the first computer-implemented system in response to the first API request, wherein the first result data comprises a first time-stamped event associated with a first device registered in association with the account identifier;
receiving second result data from the first computer-implemented system in response to the first API request, wherein the second result data comprises a second time-stamped event associated with a second device registered in association with the account identifier;
generating second prompt data comprising the first natural language input, first text data describing the first time-stamped event of the first result data, and second text data describing the second time-stamped event of the second result data;
and
generating, by the LLM using the first prompt data and the second prompt data, a first output action responsive to the first natural language input.

5 more elements

A method comprising:

receiving a first natural language input from a first device;
generating first prompt data comprising the first natural language input;
generating, by a large language model (LLM) using the first prompt data, a first directive to send a first request to a first interface of a first computer-implemented system, wherein the first computer-implemented system is different from the first device;
receiving first result data from the first computer-implemented system in response to the first request, wherein the first result data comprises a first time-stamped event associated with the first device or a second device;
generating second prompt data comprising the first natural language input and a representation of the first result data;
and
generating, by the LLM using the second prompt data, a first output action responsive to the first natural language input.

4 more elements

A system comprising:

at least one processor;
and
non-transitory computer-readable memory storing instructions that, when executed by the at least one processor, are effective to:
receive a first natural language input from a first device;
generate first prompt data comprising the first natural language input;
generate, by a large language model (LLM) using the first prompt data, a first directive to send a first request to a first interface of a first computer-implemented system, wherein the first computer-implemented system is different from the first device;
receive first result data from the first computer-implemented system in response to the first request, wherein the first result data comprises a first time-stamped event associated with the first device or a second device;
generate second prompt data comprising the first natural language input and a representation of the first result data;
and
generate, by the LLM using the second prompt data, a first output action responsive to the first natural language input.

7 more elements

Dependencies are inferred from the claim text.