US 2023/0409622 A1Published application

Determining topics for taxonomies

AssigneePinterest, Inc.

InventorsAbhijit Mahabal, Rui Huang

Life of this patent

Estimating the term…

Loading patent…

Abstract

Systems and methods for determining one or more topics that may be associated and/or grounded with a node of a taxonomy to facilitate the creation and/or modification of the taxonomy. The one or more topics can be determined by generating two layers of associations and/or groundings. In a first layer, tokens can be associated and/or grounded in a corpus of queries, and in a second layer, topics can be associated and/or grounded in the tokens. The topics can then be associated with and/or grounded in nodes of a taxonomy which can facilitate access to content items stored and maintained by an online service. Further, in exemplary implementations where the content items include associations and/or mappings to the corpus of queries, the nodes of the taxonomy (which are associated with one or more topics) can be transitively mapped to the content items.

Claims20

Dependent claims

A computer-implemented method, comprising:

for each query session of a plurality of query sessions:
determining a first plurality of tokens based at least in part on a plurality of n-grams included in a plurality of queries included in the query session;
aggregating queries included in the plurality of query sessions to generate a weighted query and token bi-graph;
generating, based at least in part on the weighted query and token bi-graph, a weighted token graph;
determining, based at least in part on the weighted token graph, a plurality of seed clusters within the weighted token graph, each of the plurality of seed clusters including a corresponding second plurality of tokens;
generating a plurality of grown seed clusters by expanding at least one of the plurality of seed clusters to include at least one additional token;
determining that a first expanded seed cluster of the plurality of seed clusters substantially overlaps a second expanded seed cluster of the plurality of expanded seed clusters;
discarding at least one of the first expanded seed cluster or the second expanded seed cluster from the plurality of expanded seed clusters to generate a reduced plurality of expanded seed clusters;
and
providing the reduced plurality of expanded seed clusters as topics.

7 more elements

A computing system, comprising:

one or more processors;
a memory storing program instructions that, when executed by the one or more processors, cause the one or more processors to at least:
determine a first plurality of tokens from a first plurality of n-grams appearing in a first plurality of queries of a first query session;
determine a second plurality of tokens from a second plurality of n-grams appearing in a second plurality of queries of a second query session;
aggregate the first plurality of queries, the second plurality of queries, the first plurality of tokens, and the second plurality of tokens to generate an aggregated query to token graph associating queries included in the first plurality of queries and the second plurality of queries to tokens included in the first plurality of tokens and the second plurality of tokens;
generate, based at least in part on the aggregated query to token graph, a token graph;
determine, based at least in part on the token graph, a plurality of seed clusters within the token graph, each of the plurality of seed clusters including a corresponding plurality of tokens;
and
provide, a first seed cluster of the plurality of seed clusters as a first topic represented by the corresponding plurality of tokens of the first seed cluster.

6 more elements

A method, comprising:

determining a plurality of tokens based on a plurality of n-grams included in queries corresponding to a plurality of query sessions;
aggregating the queries included in the plurality of query sessions to generate a weighted query and token bi-graph associating the queries to the plurality of tokens;
generating, based at least in part on the weighted query and token bi-graph, a weighted token graph;
determining, based at least in part on the weighted token graph, a plurality of seed clusters within the weighted token graph, each of the plurality of seed clusters including a corresponding plurality of tokens;
providing a first seed cluster of the plurality of seed clusters as a first topic represented by the corresponding plurality of tokens of the first seed cluster;
associating the first topic to a first node of a taxonomy;
and
mapping a plurality of content items from a corpus of content items to the first node based at least in part on the first topic.

5 more elements

Dependencies are inferred from the claim text.