IGLU: Efficient GCN Training via Lazy Updates
Authors: S Deepak Narayanan, Aditya Sinha, Prateek Jain, Purushottam Kar, SUNDARARAJAN SELLAMANICKAM
ICLR 2022 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Benchmark experiments show that IGLU offers up to 1.2% better accuracy despite requiring up to 88% less compute. Section 4, 'Empirical Evaluation', details datasets, baselines, and results, including 'Test accuracies are reported in Table 1 and convergence plots are shown in Figure 2', confirming empirical studies. |
| Researcher Affiliation | Collaboration | S Deepak Narayanan & Aditya Sinha Microsoft Research India {sdeepaknarayanan1,adityaasinha28}@gmail.com, Prateek Jain Microsoft Research India prajain@google.com, Purushottam Kar IIT Kanpur & Microsoft Research India purushot@cse.iitk.ac.in, Sundararajan Sellamanickam Microsoft Research India ssrajan@microsoft.com |
| Pseudocode | Yes | Algorithm 1 IGLU: backprop order and Algorithm 2 IGLU: inverted order |
| Open Source Code | Yes | An implementation of IGLU can be found at the following URL https://github.com/sdeepaknarayanan/iglu |
| Open Datasets | Yes | The following five benchmark tasks were used: (1) Reddit (Hamilton et al., 2017): (2) PPI-Large (Hamilton et al., 2017): (3) Flickr (Zeng et al., 2020): (4) OGBN-Arxiv (Hu et al., 2020): and (5) OGBN-Proteins (Hu et al., 2020): |
| Dataset Splits | Yes | Training-validation-test splits and metrics were used in a manner consistent with the original release of the datasets: specifically ROC-AUC was used for OGBN-Proteins and micro-F1 for all other datasets. Dataset descriptions and statistics are presented in Appendix B. Table 8 provides details on the benchmark node classification datasets used in the experiments. Train/Val/Test PPI-Large 0.79/0.11/0/10 Reddit 0.66/0.10/0.24 Flickr 0.5/0.25/0.25 OGBN-Proteins 0.65/0.16/0.19 OGBN-Arxiv 0.54/0.18/0.28 |
| Hardware Specification | Yes | We implement IGLU in Tensor Flow 1.15.2 and perform all experiments on an NVIDIA V100 GPU (32 GB Memory) and Intel Xeon CPU processor (2.6 GHz). |
| Software Dependencies | Yes | We implement IGLU in Tensor Flow 1.15.2 and perform all experiments on an NVIDIA V100 GPU (32 GB Memory) and Intel Xeon CPU processor (2.6 GHz). |
| Experiment Setup | Yes | Model Selection and Hyperparameter Tuning. Model selection was done for all methods based on their validation set performance. For IGLU, Graph SAGE and VR-GCN, an exhaustive grid search was done over general hyperparameters such as batch size, learning rate and dropout rate (Srivastava et al., 2014). In addition, method-specific hyperparameter sweeps were also carried out that are detailed in Appendix A.4. IGLU : Learning Rate {0.01, 0.001} with learning rate decay schemes, Batch Size {512, 2048, 4096, 10000}, Dropout {0.0, 0.2, 0.5, 0.7} |