Timezone: »

Bayesian Graph Neural Networks with Adaptive Connection Sampling
Arman Hasanzadeh · Ehsan Hajiramezanali · Shahin Boluki · Mingyuan Zhou · Nick Duffield · Krishna Narayanan · Xiaoning Qian

Tue Jul 14 09:00 AM -- 09:45 AM & Tue Jul 14 08:00 PM -- 08:45 PM (PDT) @

We propose a unified framework for adaptive connection sampling in graph neural networks (GNNs) that generalizes existing stochastic regularization methods for training GNNs. The proposed framework not only alleviates over-smoothing and over-fitting tendencies of deep GNNs, but also enables learning with uncertainty in graph analytic tasks with GNNs. Instead of using fixed sampling rates or hand-tuning themas model hyperparameters in existing stochastic regularization methods, our adaptive connection sampling can be trained jointly with GNN model parameters in both global and local fashions. GNN training with adaptive connection sampling is shown to be mathematically equivalent to an efficient approximation of training BayesianGNNs. Experimental results with ablation studies on benchmark datasets validate that adaptively learning the sampling rate given graph training data is the key to boost the performance of GNNs in semi-supervised node classification, less prone to over-smoothing and over-fitting with more robust prediction.

Author Information

Arman Hasanzadeh (Texas A&M University)

I am a machine learning researcher and a PhD student at Texas A&M University. My main areas of interest are learning with graph structured data, generative models, representation learning, Bayesian inference, and deep learning. My works have been applied to a variety of applications including graph analytics (e.g. network embedding, node classification, (dynamic) link prediction), drug generation, text generation, time series analysis, data integration, and anomaly detection. I have multiple publications in top-tier conferences and journals such as NeurIPS, ICML, ICASSP (best paper finalist), ISMB, and Bioinformatics.

Ehsan Hajiramezanali (Texas A&M University)
Shahin Boluki (Texas A&M University)
Mingyuan Zhou (University of Texas at Austin)
Nick Duffield (Texas A&M University)
Krishna Narayanan (Texas A&M University)
Xiaoning Qian (Texas A&M University)

More from the Same Authors