Research explainers on graph neural networks — GCNs, GATs, message passing, and graph-based representation learning.
Most of what the research community knows about backdoor attacks on graphs comes from homogeneous graphs, where every node and every edge is the same type. In that setting, an attacker can attach almost any subgraph to almost any node,…
Backdoor Attacks Now Target Heterogeneous Graph Neural Networks Read More »
Most real recommendation catalogs are lopsided by design. A small number of items collect the overwhelming majority of clicks, purchases, or plays, while a much larger number sit with only a handful of interactions each. Collaborative filtering, the workhorse behind…
CHAKG Lets Popular Items Rescue Long Tail Recommendations Read More »
Retail demand planning would be simple if every product existed in isolation. It does not. A grocery chain needs a forecast for total daily revenue, for each brand it stocks, for each store, and for each item on the shelf,…
Graph Neural Networks Bring Coherent Forecasts to Retail Read More »
AngioGraphCAD models coronary arteries as graphs, letting a graph neural network assess lesion-level risk the way cardiologists reason — one lesion at a time.
Jia He and Maggie X. Cheng from Illinois Institute of Technology have found a way to build GNNs with the same k-WL and k-FWL expressive power as the best known high-order networks — using a fraction of the parameters and…
Orthogonal Bases for Equivariant Graph Learning with Provable k-WL Expressive Power Read More »
A team from Peking University and Georgia Tech has built a formal theory explaining why the most widely used GNN approach to multi-node tasks breaks down — and proved that a simple but principled labeling strategy solves the problem completely.
Improving Graph Neural Networks on Multi-node Tasks with the Labeling Trick Read More »
A multi-institution team spanning Peking University, UC San Diego, Stony Brook University, and Weill Cornell Medicine built PEGN — a framework that injects localized topological features from extended persistent homology into graph neural networks, pushing expressiveness beyond the 3-WL barrier…
PEGN: How Persistent Homology Breaks the WL Barrier in Graph Neural Networks Read More »
A team at Chongqing University built a graph neural network that embeds multi-sensor degradation data into hyperbolic space — where the geometry naturally fits hierarchical structures — and dynamically adapts the curvature of that space during training to match the…
DCPGCN: Dynamic Curvature Pooling in Hyperbolic Space for Multi-Sensor RUL Prediction Read More »
Shan Zhao, Ioannis Prapas, and their colleagues at TU Munich and the National Observatory of Athens built a causally informed graph neural network that forecasts burned areas with 64% lower prediction variance and outperforms correlation-based baselines by 2–5 AUROC points…
Causal Graph Neural Networks for Wildfire Forecasting Across Geographic Shifts Read More »
A novel training paradigm that integrates adversarial perturbations with instance-sensitive optimization to enhance robustness and generality in graph neural network-based collaborative filtering, achieving up to 9.2% NDCG improvement over state-of-the-art methods.
ACCF: Adversarial Contrastive Collaborative Filtering Read More »