Bottom-up clustering improves semantic ids for better product search
Exploring Bottom-Up Clustering for Creating Semantic IDs
Information RetrievalArtificial Intelligence
Summary
Finding unique labels that capture the meaning of products is important for search systems that generate answers. Previous methods struggled to link each unique label to one product while keeping useful information for later tasks. This paper shows that using a bottom-up clustering approach helps organize these labels better by preserving the natural groupings of products. This leads to more meaningful and unique identifiers, which can improve how search systems find and use product information. Overall, the authors suggest that their method makes the labels more accurate and useful without losing important details.
Semantic IDgenerative retrievalembedding spacebottom-up clusteringresidual quantizationhierarchical clusteringunique identifierproduct embeddingscodewordsclustering quality
Authors
Leah Woldemariam, Sudhanshu Garg, Taha Belkhouja, Charles Kim-Yip, Ali Sahami
Abstract
The success of generative retrieval has largely been attributed to the use of Semantic IDs, which improve over arbitrary item-level identifiers such as hashes by capturing the semantics of items. The main challenges faced when constructing Semantic IDs, however, is in mapping each identifier to a unique product and capturing information valuable to downstream tasks. Past works have appended additional codewords to de-duplicate item identifiers and utilized residual quantization to create hierarchical clusters. In this work, we present an algorithm for generating Semantic IDs that ensure the identifiers are both unique and preserve the structure of the original embedding. Key to our work is the use of bottom-up clustering to preserve local structure in the embedding space, improving the clustering quality of the resulting Semantic IDs and their utility for downstream generative retrieval.