Distributed Composite Quantization

Authors: Weixiang Hong, Jingjing Meng, Junsong Yuan

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

Reproducibility Variable Result LLM Response
Research Type Experimental Extensive experiments on ANN search and image retrieval tasks validate that the proposed DCQ significantly improves Composite Quantization in both efficiency and scale, while still maintaining competitive accuracy.
Researcher Affiliation Academia Weixiang Hong, Jingjing Meng, Junsong Yuan School of Electrical and Electronic Engineering, Nanyang Technological University, Singapore {wxhong,jingjing.meng,JSYUAN}@ntu.edu.sg
Pseudocode No The paper describes the optimization process mathematically and textually but does not include a clearly labeled "Pseudocode" or "Algorithm" block.
Open Source Code No The paper states: "We use the publicly available implementation of L-BFGS 1. 1http://www.chokkan.org/software/liblbfgs/". This refers to a third-party library used, not the authors' own implementation code for their proposed method.
Open Datasets Yes We perform the ANN search experiments on three datasets: MNIST (Le Cun et al. 1998), ...; Label Me22K (Russell et al. 2008), ...; and SIFT1M (Jegou, Douze, and Schmid 2011),...
Dataset Splits No The paper mentions training data and test data but does not explicitly describe training/test/validation splits with percentages, counts, or references to predefined splits for reproducibility beyond stating "100K learning vectors and 10K queries" for SIFT1M dataset, which implies a training/test split, but not specifically validation, and not for all datasets.
Hardware Specification Yes Our machine is equipped with 24 Intel Xeon CPUs E5-2630 (2.30GHz) and 96 GB memory. ... We gratefully acknowledge the support of NVAITC (NVIDIA AI Technology Centre) for their donation of a Tesla K80 and M60 GPU used for our research at the ROSE Lab.
Software Dependencies No The paper mentions using "LBFGS algorithm (Nocedal 1980; Liu and Nocedal 1989)" and "publicly available implementation of L-BFGS", but it does not specify any version numbers for this or other software components.
Experiment Setup Yes We empirically set the penalty parameter ρ = 100 and the number of ADMM iterations I = 5.