Training-Free Binary Projection Filtering for Dense Retrieval: An Empirical Study of Candidate Reduction, Ranking Stability, and Failure Risk

Dense retrieval pipelines often rely on large candidate pools before reranking, making candidate generation and downstream scoring a practical bottleneck. This paper studies training-free binary projection filtering as a lightweight pre-filter for reducing the candidate set before dense reranking. R...

Popoln opis

Bibliografske podrobnosti

Principais autores:	Tip-aroon Kiawkaew, Thanaruk Theeramunkong
Format:	Artigo
Jezik:	Inglês
Izdano:	MDPI AG 2026-05-01
Serija:	Information
Teme:	dense retrieval candidate reduction binary projection filtering dense reranking catastrophic failure retrieval robustness
Online dostop:	https://www.mdpi.com/2078-2489/17/5/514
Oznake:	Označite Brez oznak, prvi označite!