Learning Latent Superstructures in Variational Autoencoders for Deep Multidimensional Clustering

Authors: Xiaopeng Li, Zhourong Chen, Leonard K. M. Poon, Nevin L. Zhang

ICLR 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental In Section 4, we present the empirical results. The rest of the paper is organized as follows. The related works are reviewed in Section 2. We introduce the proposed method and learning algorithms in Section 3. In Section 4, we present the empirical results. The conclusion is given in Section 5.
Researcher Affiliation Academia Xiaopeng Li1, Zhourong Chen1, Leonard K. M. Poon2 and Nevin L. Zhang1 1 Department of Computer Science and Engineering The Hong Kong University of Science and Technology 2 Department of Mathematics & Information Technology The Education University of Hong Kong {xlibo,zchenbb,lzhang}@cse.ust.hk, kmpoon@eduhk.hk
Pseudocode Yes Algorithm 1 Learning Latent Tree Variational Autoencoder
Open Source Code No The paper does not contain any explicit statement about open-source code release or provide a link to a code repository for the described methodology.
Open Datasets Yes The datasets include MNIST, STL-10, Reuters (Xie et al., 2016; Jiang et al., 2017) and the Heterogeneity Human Activity Recognition (HHAR) dataset.
Dataset Splits No The paper does not provide specific details on training, validation, and test dataset splits (e.g., percentages or exact counts), nor does it reference predefined splits with citations for these specific datasets.
Hardware Specification No The paper does not provide any specific hardware details (e.g., GPU/CPU models, memory specifications) used for running the experiments.
Software Dependencies No The paper mentions software components like 'Adam optimizer' but does not specify version numbers for any libraries, frameworks, or development environments used for the experiments.
Experiment Setup Yes For the proposed LTVAE, we use Adam optimzer (Kingma & Ba, 2015) with initial learning rate of 0.001 and mini-batch size of 128. For Stepwise EM, we set the learning rate to be 0.01. As in Algorithm 1, we set E = 5, i.e. we update the latent tree model every 5 epochs. When optimizing the candidate models during structure search, we perform 10 random restarts and train with EM for 200 iterations.