The Fuse XORier Lookup Table: Exploration, Implementation, and Revision of Probabilistic Sets and Maps

Abstract

This paper presents an exploration, implementations, and revisions of probabilistic sets and maps, specifically focusing on Bloomier filters and related data structures. The paper introduces the Fuse XORier Lookup Table (FXLT), an enhanced version of the Bloomier Filter incorporating spatial coupling, linear construction, and optimizations. The authors provide implementations in C and Python, comparing the FXLT's performance with other data structures like bloom filters, XOR filters, binary fuse filters, hash tables, and red-black trees. The FXLT demonstrates improvements in both space and time efficiency over traditional Bloomier Filters and appears competitive with hash tables for large datasets.