Non-metric Similarity Graphs for Maximum Inner Product Search

Authors: Stanislav Morozov, Artem Babenko

NeurIPS 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental By an extensive comparison to the existing approaches, we show that the proposed method is a game-changer in terms of the runtime/accuracy trade-off for the MIPS problem. We introduce a new large-scale dataset for the MIPS algorithms evaluation to facilitate research in this direction.
Researcher Affiliation Collaboration Stanislav Morozov Yandex, Lomonosov Moscow State University stanis-morozov@yandex.ru Artem Babenko Yandex, National Research University Higher School of Economics artem.babenko@phystech.edu
Pseudocode Yes Algorithm 1 Greedy walk; Algorithm 2 NSW graph construction
Open Source Code Yes The dataset and the C++ implementation of our method are available online1. Footnote 1: https://github.com/stanis-morozov/ip-nsw
Open Datasets Yes Datasets. We summarize the information on the benchmark datasets in Table 1. The Netflix, Movie Lens and Yahoo!Music datasets are the established benchmarks for the MIPS problem. Music100 is a new dataset that we introduce to the community2. This dataset was obtained by IALSfactorization[26]... Footnote 2: https://github.com/stanis-morozov/ip-nsw
Dataset Splits No The paper mentions training, testing, and evaluation metrics but does not provide specific details on validation dataset splits, percentages, or methods for creating them.
Hardware Specification Yes All the experiments were performed on Intel Xeon E5-2650 machine in a single thread mode.
Software Dependencies No The paper mentions using the 'Intel MKL library' and provides a 'C++ implementation' but does not specify version numbers for any software dependencies or libraries.
Experiment Setup Yes The primary parameter of the NSW is the maximum vertex degree M... In our experiments we use M = 32 edges per vertex. We performed the experiments with K = 1 and K = 10. The recall values were averaged over 10, 000 randomly sampled queries.