Deep Discrete Prototype Multilabel Learning

Authors: Xiaobo Shen, Weiwei Liu, Yong Luo, Yew-Soon Ong, Ivor W. Tsang

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

Reproducibility Variable Result LLM Response
Research Type Experimental Extensive experiments on several large-scale datasets demonstrate that DBPC achieves several orders of magnitude lower storage and prediction complexity than state-of-the-art multi-label methods, while achieving competitive accuracy. and 4 Experiments In this section, we evaluate the performance of the proposed method for multi-label classification.
Researcher Affiliation Collaboration Rolls-Royce@NTU Corporate Lab, Nanyang Technological University School of Computer Science and Engineering, The University of New South Wales School of Computer Science and Engineering, Nanyang Technological University Centre for Artificial Intelligence, FEIT, University of Technology Sydney
Pseudocode Yes Algorithm 1 Binary Prototype KNN Prediction and Algorithm 2 Deep Binary Prototype Compression
Open Source Code No The paper does not provide any explicit statement or link indicating the availability of the source code for the described methodology.
Open Datasets Yes DELICIOUS1 : contains textual data of web pages along with 983 tags extracted from the del.icio.us social book marking site. 1http://mulan.sourceforge.net; MIRFLICKR25K2: consists of 25,000 images collected from the social photography website Flickr. 2http://press.liacs.nl/mirflickr/; NUS-WIDE3: consists of 269,648 images from 81 ground-truth concepts with a total number of 5,018 unique tags. 3http://lms.comp.nus.edu.sg/research/NUS-WIDE.htm
Dataset Splits Yes 3-fold cross-validation is applied to split training and testing sets. and the k in k NN search is determined using 5-fold cross validation over the range {1, 5, 10, 20} for all k NN-based methods.
Hardware Specification Yes All the experiments are performed on a Ubuntu 64-Bit Linux workstation with 24core Intel Xeon CPU E5-2620 2.10 GHz and 128 GB memory.
Software Dependencies No The paper mentions 'LIBLINEAR' and 'CNN-F model' but does not provide specific version numbers for any software dependencies.
Experiment Setup Yes The dimension of the embedding in SLEEC and DBPC is set to 128 for all the datasets. ... Both encoding and decoding NNs in our proposed DBPC consist two fully-connected layers on DELICIOUS and EUR-LEX datasets, where the first layer has 4,096 nodes. ... The activation functions for the previous and last layers are Re LU and identity functions, respectively. We fix the mini-batch size to 128 and tune the learning rate from 10 6 to 10 2 by cross validation; µ and α are set to 0.1 and 0.01 respectively; τ is selected from [1, 10] by cross validation. We set the size of prototype set in DBPC as n ρ , where ratio ρ is set to 0.1 in this work.