GAIPS: Accelerating Maximum Inner Product Search with GPU

Long, Xiang; Xiao, Yan; Lu, Lan; Tang, Bo

doi:10.1145/3404835.3462997

Cited by 2 publications

(3 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This naturally gives rise to the Maximum Inner Product Search (MIPS) problem, which finds the vector in a set of n item vectors P ⊂ R d that has the largest inner product with a query (user) vector q ∈ R d , i.e., p * = arg max p∈P ⟨p, q⟩, as well as its extension kMIPS that finds k (k > 1) vectors with the largest inner products for recommending items to users. Due to its prominence in recommender systems, the kMIPS problem has attracted significant research interests, and numerous methods have been proposed to improve the search performance (Ram and Gray 2012;Koenigstein, Ram, and Shavitt 2012;Keivani, Sinha, and Ram 2018;Teflioudi and Gemulla 2017;Li et al 2017;Abuzaid et al 2019;Shrivastava and Li 2014;Neyshabur and Srebro 2015;Shrivastava and Li 2015;Huang et al 2018;Yan et al 2018;Ballard et al 2015;Yu et al 2017;Ding, Yu, and Hsieh 2019;Lorenzen and Pham 2020;Pham 2021;Shen et al 2015;Guo et al 2016;Dai et al 2020;Xiang et al 2021;Morozov and Babenko 2018;Tan et al 2019;Zhou et al 2019;Liu et al 2020;Tan et al 2021).…”

Section: Introductionmentioning

confidence: 99%

SAH: Shifting-Aware Asymmetric Hashing for Reverse k Maximum Inner Product Search

Huang

Wang

Tung

2023

AAAI

View full text Add to dashboard Cite

This paper investigates a new yet challenging problem called Reverse k-Maximum Inner Product Search (RkMIPS). Given a query (item) vector, a set of item vectors, and a set of user vectors, the problem of RkMIPS aims to find a set of user vectors whose inner products with the query vector are one of the k largest among the query and item vectors. We propose the first subquadratic-time algorithm, i.e., Shifting-aware Asymmetric Hashing (SAH), to tackle the RkMIPS problem. To speed up the Maximum Inner Product Search (MIPS) on item vectors, we design a shifting-invariant asymmetric transformation and develop a novel sublinear-time Shifting-Aware Asymmetric Locality Sensitive Hashing (SA-ALSH) scheme. Furthermore, we devise a new blocking strategy based on the Cone-Tree to effectively prune user vectors (in a batch). We prove that SAH achieves a theoretical guarantee for solving the RMIPS problem. Experimental results on five real-world datasets show that SAH runs 4~8x faster than the state-of-the-art methods for RkMIPS while achieving F1-scores of over 90%. The code is available at https://github.com/HuangQiang/SAH.

show abstract

Section: Introductionmentioning

confidence: 99%

SAH: Shifting-Aware Asymmetric Hashing for Reverse k Maximum Inner Product Search

Huang

Wang

Tung

2023

AAAI

View full text Add to dashboard Cite

show abstract

“…The inner product search (IPS) problem is important in many fields, e.g., information retrieval [18,32,35], recommender systems [6,7,30], data mining [19,27], databases [22,24,29], artificial intelligence [13,36], and machine learning [14,23,34]. These fields usually consider the 𝑘 maximum inner product search (𝑘-MIPS) problem, which, given a query vector and an output size 𝑘, returns the 𝑘 vectors having the maximum inner product with the query.…”

Section: Introductionmentioning

confidence: 99%

“…Trivially, approximation algorithms which do not guarantee that all vectors such that x • q ≥ 𝜏 are included in the search result cannot solve the FI-IPS problem correctly. This paper does not assume GPU settings[31,32], because parallelization does not overcome the essential drawback of 𝑂 (𝑛) time.…”

mentioning

confidence: 99%

Simpler is Much Faster: Fair and Independent Inner Product Search

Aoyama

Amagata

Fujita

et al. 2023

Proceedings of the 46th International ACM SIGIR Conference on Research and Development in Information Retrieval

View full text Add to dashboard Cite

The problem of inner product search (IPS) is important in many fields. Although maximum inner product search (MIPS) is often considered, its result is usually skewed and static. Users are hence hard to obtain diverse and/or new items by using the MIPS problem. Motivated by this, we formulate a new problem, namely the fair and independent IPS problem. Given a query, a threshold, and an output size 𝑘, this problem randomly samples 𝑘 items from a set of items such that the inner product of the query and item is not less than the threshold. For each item that satisfies the threshold, this problem is fair, because the probability that such an item is outputted is equal to that for each other item. This fairness can yield diversity and novelty, but this problem faces a computational challenge. Some existing (M)IPS techniques can be employed in this problem, but they require 𝑂 (𝑛) or 𝑜 (𝑛) time, where 𝑛 is the dataset size. To scale well to large datasets, we propose a simple yet efficient algorithm that runs in 𝑂 (log 𝑛 + 𝑘) expected time. We conduct experiments using real datasets, and the results demonstrate that our algorithm is up to 330 times faster than baselines. CCS CONCEPTS• Information systems → Retrieval efficiency.

show abstract

GAIPS: Accelerating Maximum Inner Product Search with GPU

Cited by 2 publications

References 11 publications

SAH: Shifting-Aware Asymmetric Hashing for Reverse k Maximum Inner Product Search

SAH: Shifting-Aware Asymmetric Hashing for Reverse k Maximum Inner Product Search

Simpler is Much Faster: Fair and Independent Inner Product Search

Contact Info

Product

Resources

About