Clustering-friendly Representation Learning via Instance Discrimination and Feature Decorrelation

Authors: Yaling Tao, Kentaro Takagi, Kouta Nakata

ICLR 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Through detailed experiments and examination, we show that the approach can be adapted to learning a latent space for clustering. In evaluations of image clustering using CIFAR-10 and Image Net-10, our method achieves accuracy of 81.5% and 95.4%, respectively.
Researcher Affiliation Industry Yaling Tao, Kentaro Takagi & Kouta Nakata Corporate R&D Center, Toshiba Corporation 1, Komukai Toshiba-cho, Saiwai-ku, Kawasaki, Kanagawa, Japan {yaling1.tao,kentaro1.takagi,kouta.nakata}@toshiba.co.jp
Pseudocode No The paper does not include pseudocode or clearly labeled algorithm blocks; the methods are described in text and mathematical formulas.
Open Source Code Yes Our Py Torch Paszke et al. (2019) implementation of IDFD is available at https://github.com/TTN-YKK/Clustering_ friendly_representation_learning.
Open Datasets Yes We conducted experiments using five datasets: CIFAR-10 Krizhevsky et al. (2009), CIFAR-100 Krizhevsky et al. (2009), STL-10 Coates et al. (2011), Image Net-10 Deng et al. (2009), and Image Net-Dog Deng et al. (2009).
Dataset Splits No The paper mentions "Specifically, the training and testing sets of dataset STL-10 were jointly used in our experiments," but does not provide explicit training/validation/test splits with percentages, sample counts, or clear references to predefined splits for all datasets, nor does it specify a separate validation set.
Hardware Specification No The paper does not provide specific hardware details such as GPU or CPU models, memory, or other detailed computer specifications used for running the experiments.
Software Dependencies No The paper mentions using "Py Torch" but does not specify a version number or list other software dependencies with their corresponding version numbers.
Experiment Setup Yes The dimension of latent feature vectors was set to d = 128, and a stochastic gradient descent optimizer with momentum β = 0.9 was used. The learning rate lr was initialized to 0.03, then gradually scaled down after the first 600 epochs using a coefficient of 0.1 every 350 epochs. The total number of epochs was set to 2000, and the batch size was set to B = 128. Orthogonality constraint weights for IDFO were α = 10 for CIFAR-10 and CIFAR-100 and α = 0.5 for the STL-10 and Image Net subsets. The weight for IDFO α was set according to the orders of magnitudes of the two losses LI and LF O. For IDFD, the weight α was simply fixed at 1. In the main experiments, we set the default temperature parameter value τ = 1 for ID(tuned), IDFO, and IDFD, and τ2 = 2 for IDFD.