Diffusion-Convolutional Neural Networks

Authors: James Atwood, Don Towsley

NeurIPS 2016 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental In this section we present several experiments to investigate how well DCNNs perform at node and graph classification tasks. In each case we compare DCNNs to other well-known and effective approaches to the task. Table 1: A comparison of the performance between baseline ℓ1 and ℓ2-regularized logistic regression models, exponential diffusion and Laplacian exponential diffusion kernel models, loopy belief propagation (LBP) on a partially-observed conditional random field (CRF), and a two-hop DCNN on the Cora and Pubmed datasets.
Researcher Affiliation Academia James Atwood and Don Towsley College of Information and Computer Science University of Massachusetts Amherst, MA, 01003 {jatwood|towsley}@cs.umass.edu
Pseudocode No No pseudocode or algorithm blocks are provided.
Open Source Code No The paper states: 'The model was implemented in Python using Lasagne and Theano [3]', but does not provide concrete access (link, explicit statement of release) to the source code for the DCNN model itself.
Open Datasets Yes The graphs were constructed from the Cora and Pubmed datasets, which each consist of scientific papers (nodes), citations between papers (edges), and subjects (labels). ... The Cora corpus [5]... The Pubmed corpus [5]... We apply DCNNs to a standard set of graph classification datasets that consists of NCI1, NCI109, MUTAG, PCI, and ENZYMES. The NCI1 and NCI109 [7] datasets... MUTAG [8]... PTC [9]... ENZYMES [10].
Dataset Splits Yes During each trial, the input graph s nodes are randomly partitioned into training, validation, and test sets, with each set having the same number of nodes. In this experiment, the validation and test set each contain 10% of the nodes, and the amount of training data is varied between 10% and 100% of the remaining nodes. At the beginning of each trial, input graphs are randomly assigned to training, validation, or test, with each set having the same number of graphs.
Hardware Specification No The paper states 'efficiently implemented on a GPU' and mentions 'NVIDIA through the donation of equipment used to perform experiments', but does not provide specific hardware details such as GPU models, CPU types, or memory specifications.
Software Dependencies No The paper mentions 'The model was implemented in Python using Lasagne and Theano [3]', but does not provide specific version numbers for these software components.
Experiment Setup Yes In each of the following experiments, we use the Ada Grad algorithm [2] for gradient ascent with a learning rate of 0.05. All weights are initialized by sampling from a normal distribution with mean zero and variance 0.01. We choose the hyperbolic tangent for the nonlinear differentiable function f and use the multiclass hinge loss between the model predictions and ground truth as the training objective.