Graph-Based Machine Learning Techniques

Dear Colleagues,

We invite submissions to a Special Issue on Graph-Based Machine Learning (GML), focusing on advances that leverage relational structures to learn from complex, interconnected data. Graphs pervade modern data ecosystems—knowledge graphs, molecular and protein interaction networks, transportation and power grids, and social and financial networks—where performance hinges on modeling entities and their relationships. Recent progress in graph neural networks and related methods has unlocked powerful capabilities for reasoning, prediction, and generation, yet key challenges remain in scalability, heterogeneity, dynamics, reliability, and responsible deployment. This Special Issue aims to consolidate cutting-edge theory, methods, systems, and applications that push GML toward practical, trustworthy, and impactful use.

As AI systems move beyond i.i.d. samples to structured, multi-relational settings, GML provides the inductive biases and operators needed for generalization under distribution shifts, compositional reasoning, and multi-hop inference. The rapid convergence of GML with foundation models (e.g., graph-augmented RAG, neurosymbolic integration, tool-using agents), the rise of dynamic and heterogeneous graphs, and the demand for transparent, fair, robust, and private models underscore the urgency and importance of this research direction across science and industry.

Topics of interest include, but are not limited to, the following:

• Novel GNN architectures; self-/semi-supervised, contrastive, and transfer learning on graphs;
• Dynamic, temporal, heterogeneous, attributed, and hypergraph learning;
• Graph generative models (e.g., molecules, materials, program graphs) and structure-aware diffusion;
• Scalability: billion-edge graphs, streaming/incremental learning, distributed and efficient systems;
• Explainability, fairness, robustness, privacy, and security of graph models; adversarial graphs;
• Out-of-distribution generalization, causal and counterfactual reasoning on graphs;
• Foundation models with graphs: graph-augmented RAG, graph prompting, LLM–graph integration, neurosymbolic methods;
• Evaluation: benchmarks, metrics, reproducibility, compute reporting, and open resources;
• Applications in biomedicine and drug discovery, recommender systems, finance and fraud, mobility and logistics, cybersecurity, climate and energy, and scientific discovery.

Prof. Dr. Zsolt Tibor Kosztyán
Guest Editor

Manuscript Submission Information

Manuscripts should be submitted online at www.mdpi.com by registering and logging in to this website. Once you are registered, click here to go to the submission form. Manuscripts can be submitted until the deadline. All submissions that pass pre-check are peer-reviewed. Accepted papers will be published continuously in the journal (as soon as accepted) and will be listed together on the special issue website. Research articles, review articles as well as short communications are invited. For planned papers, a title and short abstract (about 100 words) can be sent to the Editorial Office for announcement on this website.

Submitted manuscripts should not have been published previously, nor be under consideration for publication elsewhere (except conference proceedings papers). All manuscripts are thoroughly refereed through a single-blind peer-review process. A guide for authors and other relevant information for submission of manuscripts is available on the Instructions for Authors page. Mathematics is an international peer-reviewed open access semimonthly journal published by MDPI.

Please visit the Instructions for Authors page before submitting a manuscript. The Article Processing Charge (APC) for publication in this open access journal is 2600 CHF (Swiss Francs). Submitted papers should be well formatted and use good English. Authors may use MDPI's English editing service prior to publication or during author revisions.

• graph machine learning
• graph neural networks (GNNs)
• message passing neural networks (MPNNs)
• graph transformers
• graph attention networks (GAT)
• geometric deep learning
• graph representation learning
• node/edge/graph classification
• link prediction
• community detection
• graph matching and graph kernels
• graph similarity search
• hypergraphs and hypergraph learning
• heterogeneous information networks
• dynamic/temporal graphs
• knowledge graphs and reasoning
• self-supervised graph learning
• contrastive learning on graphs
• transfer/few-shot learning on graphs
• pretraining for graphs
• graph generative models
• molecular and materials property prediction
• drug discovery and molecular design
• recommender systems
• financial risk and fraud detection
• cybersecurity and threat intelligence
• mobility, transportation, and traffic networks
• power grids and infrastructure networks
• scientific discovery on graphs
• out-of-distribution generalization (OOD) for graphs
• causal and counterfactual inference on graphs
• graph anomaly detection
• adversarial attacks and robustness in GML
• explainability and interpretability for graph models
• fairness, ethics, and responsible GML
• privacy-preserving and federated graph learning
• scalable/billion-scale graph learning
• distributed and efficient GML systems
• streaming/online and incremental graph learning
• graph sampling and mini-batching
• evaluation, benchmarks, and reproducibility
• open datasets, code, and model cards
• Foundation Models with Graphs
• graph-augmented RAG and graph prompting
• LLM–graph integration and tool use
• neurosymbolic and logic-guided GML
• energy-efficient and sustainable GML
• uncertainty estimation and calibration on graphs
• graph databases, querying, and storage systems
• combinatorial optimization on graphs