Susheel Suresh

I am a PhD Student in the Computer Science department at Purdue University. I am advised by Prof. Jennifer Neville.

My research interests lie in the fields of machine learning and data mining. Specifically on developing learning algorithms and principles for representation learning on graph structured data.

My recent work can be broadly categorized into the following themes:

Analysis of graph neural networks with a network science lens.
Principles for self-supervised graph contrastive learning.
Temporal graph representation learning with applications to real-world tasks.

My current focus is on various interrelated aspects of machine learning for graph data viz. domain adaptation/generalization, invariant learning and adversarial robustness.

In the past, I have also worked on developing hybrid knowledge graph mining methods and one of my long standing research goals is to unify neural based inductive learning with deductive symbolic methods.

Apart from research, I am into badminton and long-distance running.

selected publications

NeurIPS

Adversarial Graph Augmentation to Improve Graph Contrastive Learning

Suresh, Susheel, Li, Pan, Hao, Cong, and Neville, Jennifer

In Thirty-fourth Conference on Neural Information Processing Systems 2021

Abstract HTML PDF Code Talk Reviews

Self-supervised learning of graph neural networks (GNN) is in great need because of the widespread label scarcity issue in real-world graph/network data. Graph contrastive learning (GCL), by training GNNs to maximize the correspondence between the representations of the same graph in its different augmented forms, may yield robust and transferable GNNs even without using labels. However, GNNs trained by traditional GCL often risk capturing redundant graph features and thus may be brittle and provide sub-par performance in downstream tasks. Here, we propose a novel principle, termed adversarial-GCL (AD-GCL), which enables GNNs to avoid capturing redundant information during the training by optimizing adversarial graph augmentation strategies used in GCL. We pair AD-GCL with theoretical explanations and design a practical instantiation based on trainable edge-dropping graph augmentation. We experimentally validate AD-GCL2 by comparing with the state-of-the-art GCL methods and achieve performance gains of up-to 14% in unsupervised, 6% in transfer, and 3% in semi-supervised learning settings overall with 18 different benchmark datasets for the tasks of molecule property regression and classification, and social network classification
KDD

Breaking the Limit of Graph Neural Networks by Improving the Assortativity of Graphs with Local Mixing Patterns

Suresh, Susheel, Budde, Vinith, Neville, Jennifer, Li, Pan, and Ma, Jianzhu

In Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery & Data Mining 2021

Abstract HTML PDF Code Talk Reviews

Graph neural networks (GNNs) have achieved tremendous success on multiple graph-based learning tasks by fusing network structure and node features. Modern GNN models are built upon iterative aggregation of neighbor’s/proximity features by message passing. Its prediction performance has been shown to be strongly bounded by assortative mixing in the graph, a key property wherein nodes with similar attributes mix/connect with each other. We observe that real world networks exhibit heterogeneous or diverse mixing patterns and the conventional global measurement of assortativity, such as global assortativity coefficient, may not be a representative statistic in quantifying this mixing. We adopt a generalized concept, node-level assortativity, one that is based at the node level to better represent the diverse patterns and accurately quantify the learnability of GNNs. We find that the prediction performance of a wide range of GNN models is highly correlated with the node level assortativity. To break this limit, in this work, we focus on transforming the input graph into a computation graph which contains both proximity and structural information as distinct type of edges. The resulted multi-relational graph has an enhanced level of assortativity and, more importantly, preserves rich information from the original graph. We then propose to run GNNs on this computation graph and show that adaptively choosing between structure and proximity leads to improved performance under diverse mixing. Empirically, we show the benefits of adopting our transformation framework for semi-supervised node classification task on a variety of real world graph learning benchmarks.