Sparse-Matrix Factorization-based Method: Efficient Computation of Latent Query and Item Representations to Approximate CE Scores

Cross-encoder (CE) models evaluate similarity by simultaneously encoding a query-item pair, outperforming the dot-product with embedding-based models at estimating query-item relevance. Current methods perform k-NN search with CE by approximating the CE similarity with a vector embedding space fit with dual-encoders (DE) or CUR matrix factorization. However, DE-based methods face challenges from poor recall because […]

The post Sparse-Matrix Factorization-based Method: Efficient Computation of Latent Query and Item Representations to Approximate CE Scores appeared first on MarkTechPost.

ai paper summary ai shorts artificial intelligence computation current editors pick embedding encoder encoding factorization matrix product query search space sparse-matrix staff tech news technology vector