Preprint
Oct 2026
FlexVDR: Learning to Allocate Matryoshka Representation Budgets for Visual Document Retrieval
FlexVDR is introduced, a framework for query-adaptive budget allocation within a single retriever and a Budget Allocator predicts route-specific retrieval utility from early query features and combines these predictions with a configurable resource preference to select a route.
Hao-Wen Xiang, Yibo Yan, Xin Zou et al.
· 0 citations