PROFESSIONAL ELECTIVES

Electives in Artificial Intelligence

Unit I

Introduction to Machine Learning for Graphs, Structure of Graphs, Node Embeddings, Random graphs with arbitrary degree distributions and their applications, Properties of Networks, and Random Graph Models, Motifs and Structural Roles in Networks, Simple Building Blocks of Complex Networks, Community Structure in Networks, Fast unfolding of communities in large networks, Overlapping Community Detection at Scale: A Nonnegative Matrix Factorization Approach, Spectral Clustering, Message Passing and Node Classification, Graph Representation Learning

Unit II

Theory of Graph Neural Networks, Architectures-GCN, GAT, MPNN & Design Space, Deep Generative Models for Graphs, Link Analysis: PageRank, Network Effects and Cascading Behaviour, Probabilistic Contagion and Models of Influence, Influence Maximization in Networks, Outbreak Detection in Networks, Network Evolution, Reasoning over Knowledge Graphs, Applications of Graph Neural Networks

Unit III

Efficient Graphlet Kernels for Large Graph Comparison, Semi-Supervised Classification with Graph Convolutional Networks, Inductive Representation Learning on Large Graphs ,Graph Attention Networks, GNN Augmentation and Training, Hierarchical Graph Representation Learning with Differentiable Pooling, Machine Learning with Heterogeneous Graphs, Modeling Relational Data with Graph Convolutional Networks, Heterogeneous Graph Transformer, Advanced Topics in GNNs, Algorithm for Training Deep and Large Graph Convolutional Networks

Course Objectives

• This course aims to equip students with a solid foundation in the theory and applications of machine learning with graphs, as well as practical skills in implementing and applying GNNs to real-world problems.
• By the end of the course, students should be able to understand and use various GNN architectures and techniques to solve graph-based machine learning problems.

Course Outcomes

CO1: Develop an understanding of the theory and applications of machine learning with graphs.

CO2: Gain an understanding of various GNN architectures and techniques.

CO3: Acquire practical skills in implementing and applying GNNs to solve real-world problems.

CO4: Be able to handle large graphs and solve graph-based machine learning problems using GNNs.

CO-PO Mapping

Evaluation Pattern: 70:30

*CAT – Can be Quizzes, Assignments, and Reports

*CAL – Can be Lab Assessments, Project, and Report

**End Semester can be theory examination/ lab-based examination/ project presentation

Textbook(s)

William L. Hamilton, “Graph Representation Learning”, McGill University 2020.

David Easley and Jon Kleinberg, “Networks, Crowds, and Markets: Reasoning About a Highly Connected World”, Cambridge University Press (2010).

Reference(s)

Negro, Alessandro. “Graph-powered machine learning”. Simon and Schuster, 2021.

Pósfai, Márton, and Albert-Laszlo Barabasi. “Network Science”. Cambridge, UK: Cambridge University Press, 2016.