Back to results
US 11,899,714 B1Granted
Voice conversational search constraint management
AssigneeAmazon Technologies, Inc.
InventorsEdward Bueche, Francois Mairesse, Amina Shabbeer, Warren D. Freitag, Jonathan Pollack, Charles Lee Thorp
Life of this patent
Estimating the term…
Abstract
Voice data from a current conversation between a user and a voice-controlled user device can be used to determine a search constraint for searching a database. Other search constraints can be determined based at least in part on the current conversation, a previous conversation, and/or a previous action. Properties can be associated with the search constraints. Once the search constraints have been determined, a plurality of search query plans is determined and a first search query plan is executed to query the database.
Claims20
A system, comprising:
- a memory configured to store computer-executable instructions;
- and
- a processor configured to access the memory and execute the computer-executable instructions to at least:
- receive, via a voice interface of a voice-controlled device, first voice data associated with a current conversation of a current search session comprising a plurality of exchanges between the voice-controlled device and a user, the current conversation relating to music;
- identify one or more concepts represented by the first voice data based at least in part on a set of concept rules;
- determine a first search constraint based at least in part on the first voice data;
- associate the first search constraint with first two or more search constraint properties based at least in part on the one or more concepts, the first two or more search constraint properties selected from a set of properties comprising a direct property representing that a respective search constraint was derived from a statement represented by the first voice data and maps directly to metadata associated with a set of electronic content items, an unselective property representing that a respective search constraint is narrowing and imprecise, a selective property representing that a respective search constraint is likely to return to search results that meet a goal of a requester, an inferred property representing that a respective search constraint was inferred from the first voice data, and a transformed property representing that a respective search constraint with a particular unselective property has been transformed to a different search constraint with a particular selective property, the first search constraint representative of at least one concept of the one or more concepts;
- provide, via the voice interface of the voice-controlled device, a request for information based at least in part on the first voice data, the request associated with the current conversation;
- responsive to the request for information, receive, via the voice interface of the voice-controlled device, second voice data associated with the current conversation;
- determine a second search constraint based at least in part on the second voice data, the second search constraint comprising second two or more search constraint properties selected from the set of properties;
- determine a plurality of search query plans for searching a database based at least in part on the first search constraint including the two or more search constraint properties, the second search constraint, and other search constraints, individual search query plans comprising unique combinations of one or more search constraints selected from the first search constraint, the second search constraint, and the other search constraints, and the database comprising a plurality of electronic content items;
- query the database by executing a first search query plan of the plurality of search query plans;
- and
- provide, via the voice interface of the voice-controlled device, information associated with one or more electronic content items of the plurality of electronic content items identified during executing the first search query plan.
11 more elements
A computer-implemented method, comprising:
- receiving voice data associated with a current conversation of a current search session comprising a plurality of exchanges between a voice-controlled device and a user;
- identifying one or more concepts represented by the voice data based at least in part on a set of concept rules;
- determining a search constraint based at least in part on the voice data, the search constraint representative of at least one concept of the one or more concepts;
- associating two or more search constraint properties with the search constraint based at least in part on the voice data and the one or more concepts, the two or more search constraint properties selected from a set of properties comprising a direct property representing that a respective search constraint was derived from a statement represented by the voice data and maps directly to metadata associated with a set of electronic content items, an unselective property representing that a respective search constraint is narrowing and imprecise, a selective property representing that a respective search constraint is likely to return to search results that meet a goal of the user, an inferred property representing that a respective search constraint was inferred from the voice data, and a transformed property representing that a respective search constraint with a particular unselective property has been transformed to a different search constraint with a particular selective property;
- determining a plurality of other search constraints based at least in part on one or more of the current conversation, a previous conversation, or a previous action;
- determining a plurality of search query plans for searching a database based at least in part on the search constraint including the two or more search constraint properties and the other search constraints, individual search query plans comprising unique combinations of search constraints selected from a set of search constraints comprising the search constraint and the other search constraints;
- and
- querying the database using one or more search query plans of the plurality of search query plans.
5 more elements
One or more non-transitory computer-readable media comprising computer-executable instructions that, when executed by one or more computer systems, cause the one or more computer systems to perform operations comprising:
- receiving, via a voice interface of a voice-controlled device, voice data associated with a current search session of a current conversation between a user and the voice-controlled device, the current conversation comprising a plurality of exchanges between the voice-controlled device and the user;
- determining whether a user goal can be derived from the voice data;
- in accordance with a determination that the user goal cannot be derived from the voice data with a sufficient level of confidence,
- identifying one or more concepts represented by the voice data based at least in part on a set of concept rules;
- determining a first search constraint based at least in part on the voice data, the first search constraint comprising two or more search constraint properties associated with the voice data and selected from a set of properties comprising a direct property representing that a respective search constraint was derived from a statement represented by the voice data and maps directly to metadata associated with a set of electronic content items, an unselective property representing that a respective search constraint is narrowing and imprecise, a selective property representing that a respective search constraint is likely to return to search results that meet a goal of a requester, an inferred property representing that a respective search constraint was inferred from the voice data, and a transformed property representing that a respective search constraint with a particular unselective property has been transformed to a different search constraint with a particular selective property, the first search constraint representative of at least one concept of the one or more concepts;
- determining a plurality of search query plans for searching a database based at least in part on the first search constraint including the two or more search constraint properties, individual search query plans comprising unique combinations of search constraints selected from a set of search constraints comprising the first search constraint and other search constraints;
- querying the database using a first search query plan of the plurality of search query plans;
- and
- providing search results for presentation at the voice-controlled device, the search results corresponding to the first search query plan.
6 more elements
Dependencies are inferred from the claim text.