Tutorials for Machine Learning on Graphs
[IJCNN 2021] Unified Spatio-Temporal modeling for traffic forecasting using Graph Convolutional Network
Research Project I completed under Dr Vinti Agrawal at BITS Pilani.
Data and code for Salesforce Research paper, GAEA: Graph Augmentation for Equitable Access via Reinforcement Learning - https://arxiv.org/abs/2012.03900 . The paper provides methods for constraint graph augmentation and optimal facility placement problems
Pure Go machine learning framework. Train, run, and serve ML models with go build. Zero CGo.
Wire-transfer AML detection with heterogeneous graph transformers on IBM's IT-AML dataset. FastAPI + D3.js monolith.
Production-grade fraud detection pipeline with entity-level behavioral feature engineering, velocity anomaly detection, graph-based risk signals, and real-time scoring API. Built with XGBoost, PyTorch, FastAPI, and Docker.
Machine learning on graphs
Graph-native AML investigation: 14 topology features, LightGBM scoring (ROC-AUC 0.87), path-level explanations, live case explorer.
Graph-RAG for Customer Journey Intelligence using NetworkX + LLM. Path-aware retrieval outperforming vector RAG on temporal queries, cohort comparison with real statistics, 5 pre-built analytics queries, and fully dockerized FastAPI/Streamlit architecture deployed on HuggingFace Spaces.
Conditional VGAE for generating synthetic temporal contact networks from node metadata
word2vec, sentence2vec, machine reading comprehension, dialog system, text classification, pretrained language model (i.e., XLNet, BERT, ELMo, GPT), sequence labeling, information retrieval, information extraction (i.e., entity, relation and event extraction), knowledge graph, text generation, network embedding
Compare LLM text embeddings with structure-aware Graph AI (GNN link prediction) on any dataset with nodes, text, and edges.
A deep learning architecture combining spectral graph neural networks with curriculum learning for HOMO-LUMO gap prediction on PCQM4Mv2. Features a dual-view architecture with Chebyshev polynomial-based spectral convolutions and complexity-driven training schedules.
Self-Supervised Similarity Learning of Floor Layouts
LaTeX research writing repo with structured build workflow and supporting artifacts.
Use NetworkViz to visualize IP Traffic flow as Graph ML problem
A deep learning approach for molecular property prediction that introduces hierarchical attention pooling to capture scaffold-aware representations. The model aggregates atom features within functional groups before global pooling, combined with scaffold-based curriculum learning for improved generalization across diverse chemical structures.
Add a description, image, and links to the graph-ml topic page so that developers can more easily learn about it.
To associate your repository with the graph-ml topic, visit your repo's landing page and select "manage topics."