作者: Gottfried Herold , Elena Kirshanova
DOI: 10.1007/978-3-662-54365-8_2
关键词: Euclidean domain 、 Computer science 、 Hash function 、 Algorithm 、 Euclidean distance 、 Euclidean shortest path 、 Space (mathematics)
摘要: We present an algorithm for the approximate k-List problem Euclidean distance that improves upon Bai-Laarhoven-Stehle (BLS) from ANTS’16. The improvement stems observation almost all solutions to form a particular configuration in n-dimensional space. Due special properties of configurations, it is much easier verify whether k-tuple forms rather than checking gives solution problem. Thus, phrasing as finding such configurations immediately better algorithm. Furthermore, search can be sped up using techniques Locality-Sensitive Hashing (LSH). Stated terms configuration-search, our LSH-like offers broader picture on previous LSH algorithms.