gnn-architectures

G3: A Programmable GNN Training System on GPU

with GUG, Let's explore the Graph Neural Network!

VISION is a framework for robust and interpretable code vulnerability detection using counterfactual data augmentation. It leverages GNNs, LLM-generated counterfactuals, and graph-based explainability to mitigate spurious correlations and improve generalization on real-world vulnerabilities (CWE-20).

A production-grade framework for predicting clinical comorbidities in EHR data using an ensemble of GNN architectures (GCN, GraphSAGE, GAT, and Graph Transformer). Features 5-dimensional clinical vectorization, multi-head attention mechanisms, and an inductive inference engine.

Deep Learning with Graph Representation of Bio-Molecules to estimate physical Properties

AI-driven generative quantum circuit design for molecular & materials discovery — GNN + Transformer + QD-GRPO reinforcement learning + CUDA-Q, validated on QPU hardware. GIC 2026.

A complete Graph Neural Network pipeline for drug molecule property prediction (BBBP). Features GCN, GAT, and GIN architectures, classical ML baselines, ablation studies, and model interpretability using GNNExplainer.

Structure-aware graph neural networks for small RNA classification.

A Python library for building AI agents that leverage the full power of Google Antigravity.

This repository contains code implementations for Graph Neural Networks (GNNs). GNNs are a category of deep learning models tailored for tasks involving graph-structured data. The provided code enables users to explore and apply GNNs for tasks such as node classification, link prediction, and graph classification.

Arquitecturas híbridas para el pronóstico de series temporales en el Distrito Metropolitano de Quito (DMQ).