Relevance-Based Embeddings for Candidate Retrieval
A paper describes query and item embeddings for efficiently retrieving candidates when relevance computation is expensive. The reported approach uses relevance to selected support items or queries.
The paper examines efficient candidate retrieval when the relevance function is computationally expensive. It describes using query and item embeddings to search for candidates without directly calculating costly relevance for every possibility.
According to the post, Yandex bases these representations on relevance to selected support items or queries. The topic may interest engineers exploring candidate retrieval with expensive relevance models. If you use AI to study or apply this approach, avoid sending personal data or internal information unless necessary, and follow your organization's data policy.