Google patents image search that learns from your clicks

1. A computing system, the system comprising: one or more processors; and one or more non-transitory computer-readable media that collectively store instructions that, when executed by the one or more processors, cause the computing system to perform operations, the operations comprising: obtaining a first image query, wherein the first image query comprises first image data; determining a first set of search results based on the first image data, wherein the first set of search results comprises a plurality of first image search results; providing the first set of search results for display in a search results interface; obtaining a selection of a particular first image search result of the plurality of first image search results, wherein the particular first image search result is associated with second image data, wherein the first image data and the second image data differ; processing the first image data with an embedding model to generate a first embedding, wherein the first embedding is associated with an embedding space; processing the second image data with the embedding model to generate a second embedding, wherein the second embedding is associated with the embedding space; determining a third embedding based on a projection prediction associated with an embedding change from the first embedding to the second embedding; determining a second set of search results based on the third embedding; and providing the second set of search results for display in the search results interface.