Course Overview:
This course equips you with the foundational knowledge of Graph Neural Networks (GNNs), a powerful AI technique specifically designed to analyze data structured as graphs. You'll explore how GNNs unlock valuable insights from interconnected information within your Supply Chain Management (SCM) processes, leading to improved decision-making and optimization.
Learning Objectives:
Define Graph Neural Networks and their core principles compared to traditional neural networks.
Understand the fundamental building blocks of GNNs (message passing, aggregation functions).
Explore different GNN architectures (e.g., GCNs, GraphSage) and their suitability for various SCM applications.
Identify real-world applications of GNNs in Supply Chain Management (e.g., demand forecasting, supplier risk assessment, network optimization).
Analyze the advantages and limitations of GNNs compared to other network analysis techniques.
Course Highlights:
1. Unveiling Graph Neural Networks
Introduction to Graphs: Representing relationships in data using nodes and edges.
Limitations of traditional neural networks for graph-structured data.
Demystifying Graph Neural Networks (GNNs): How GNNs learn from interconnected information.
Understanding message passing: The core concept of GNNs for information exchange between nodes.
Hands-on Exercises: Utilizing online tools or simple coding exercises to explore basic GNN functionalities.
Case Studies: Exploring early applications of GNNs in predicting product demand based on customer relationships within a network.
2. GNN Applications in SCM and Beyond
Exploring different GNN architectures: Graph Convolutional Networks (GCNs) and their applications in SCM.
Understanding GraphSage: A versatile GNN architecture for various network analysis tasks.
Applying GNNs to supplier risk assessment: Identifying potential disruptions within your supply network.
GNNs for network optimization: Optimizing transportation routes and logistics based on network structure.
Course Wrap-up: Addressing limitations of GNNs and responsible AI practices in SCM implementations.
Prerequisites:
Strong understanding of linear algebra, calculus, and probability theory
Proficiency in programming with Python and deep learning frameworks (e.g., TensorFlow, PyTorch)
Familiarity with basic machine learning concepts and techniques
Knowledge of graph theory and network analysis is beneficial but not required