Are Graph Neural Networks Optimal Approximation Algorithms?

Authors: Morris Yau, Nikolaos Karalias, Eric Lu, Jessica Xu, Stefanie Jegelka

NeurIPS 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Our approach achieves strong empirical results across a wide range of real-world and synthetic datasets against solvers and neural baselines.
Researcher Affiliation Academia Morris Yau MIT CSAIL morrisy@mit.edu Nikolaos Karalias MIT CSAIL stalence@mit.edu Eric Lu Harvard University ericlu01@g.harvard.edu Jessica Xu Independent researcher formerly at MIT jessica.peng.xu@gmail.com Stefanie Jegelka TUM and MIT stefje@mit.edu
Pseudocode Yes Empirically, Opt GNN is simple to implement (see pseudocode in Appendix 5) and For pseudocode, please refer to the appendix (Max-Cut: algorithm 4 and general SDPs: algorithm 5).
Open Source Code Yes Code is available at: https://github.com/penlu/bespoke-gnn4do.
Open Datasets Yes Our real-world datasets are ENZYMES, PROTEINS, MUTAG, IMDB-BINARY, COLLAB (which we will together call TU-small), and REDDIT-BINARY, REDDIT-MULTI-5K, and REDDIT-MULTI-12K (which we will call TU-REDDIT). and We also tested Opt GNN on the GSET benchmark instances (Benlic & Hao, 2013).
Dataset Splits Yes For each dataset we hold out a validation and test slice for evaluation. In our generated graph experiments we set aside 1000 graphs each for validation and testing. (...) For TU-small, we used a train/validation/test split of 0.8/0.1/0.1. For TU-REDDIT, we set aside 100 graphs each for validation and testing.
Hardware Specification Yes Our training runs used 20 cores of an Intel Xeon Gold 6248 (for data loading and random graph generation) and a NVIDIA Tesla V100 GPU. Our Gurobi runs use 8 threads on a Intel Xeon Platinum 8260. Our Ka MIS runs use an Intel Core i9-13900H. Our Lw D and DGL-TREESEARCH runs use an Intel Core i9-13900H and an RTX 4060.
Software Dependencies No Our Gurobi runs use 8 threads on a Intel Xeon Platinum 8260. and with a standard automatic differentiation package like Py Torch (Paszke et al., 2019).
Experiment Setup Yes We ran each experiment on a range of hyperparameters. See Figure 5 for the hyperparameter listing. For all training runs, we used the Adam optimizer Kingma & Ba (2014) with a learning rate of 0.001. (...) For Max-3-SAT, we set the penalty term ρ = 0.003.