What is it about?

Locality-Sensitive Hashing (LSH) and its variants are the well-known indexing schemes for the Nearest Neighbor Search (NNS) problem in high-dimensional Euclidean space. In this paper, we introduce a novel concept of query-aware bucket partition, which uses a given query as the anchor for bucket partition. With the query-aware bucket partition, we propose a novel provable Query-Aware LSH (QALSH) scheme for c-NNS over external memory, which works with an approximation ratio c > 1. Extensive experiments show that QALSH outperforms two state-of-the-art LSH schemes C2LSH and LSB-Forest, especially in high-dimensional spaces.

Featured Image

Why is it important?

Vanilla LSH functions are constructed in a query-oblivious manner in the sense that buckets are partitioned before any query arrives. However, objects closer to a query may be partitioned into different buckets, which leads to a high search error. Moreover, due to the use of query-oblivious bucket partition, the state-of-the-art LSH schemes for external memory, namely C2LSH and LSB-Forest, only work with an integer c ≥ 2. In order to fix those problems, we design a novel query-aware LSH function based on the query position. We also simply the function as a random projection coupled with the query-aware bucket partition, which removes the random shift required by traditional query-oblivious LSH functions. Notably, query-aware bucket partition can be easily implemented so that query performance is guaranteed.

Perspectives

I hope this article could bring new thought to people who design basic data structures for NNS problem or other fundamental similarity search problems. The query-aware way for NNS is quite exciting and insightful, which might be the potential of the separate interest in other computer science fields.

Dr Qiang Huang
National University of Singapore

Read the Original

This page is a summary of: Query-aware locality-sensitive hashing for approximate nearest neighbor search, Proceedings of the VLDB Endowment, September 2015, VLDB Endowment,
DOI: 10.14778/2850469.2850470.
You can read the full text:

Read

Resources

Contributors

The following have contributed to this page