Course

Introduction – Types of knowledge networks, Applications of knowledge networks, basic network metrics- nodes, edges, degree distribution, centrality measures-degree centrality, Eigen vector centrality. Network density, clustering co-efficient. Knowledge Network Dynamics – Information Diffusion and Community Evolution.

Data collection and network construction – data preprocessing, network visualization techniques, Gephi and NetworkX. Community detection – Graph Partitioning, Girvan-Newman Algorithm, Louvain Method, Latent Dirichlet Allocation (LDA), Evaluation of Community Detection: Modularity Score, Normalized Mutual Information (NMI).

Machine learning and Knowledge networks – Graph Embeddings: Node2Vec, DeepWalk, GraphSAGE. Graph Neural Networks (GNNs): Graph Attention Networks (GATs), Graph Autoencoders. Link Prediction using Unsupervised Methods: Using similarity measures (e.g., Jaccard coefficient, Adamic/Adar index). Node Classification: Semi-Supervised Learning, Label Propagation. Knowledge Graph Completion- BERT-based Models. Event Detection- Change Point Detection, Burst Detection. Case study: Connected papers and Google Knowledge Graph tools.

Pre-Requisite(s): None
Course Type: Theory

Course Objectives

• To understand the structural properties and patterns of knowledge networks using metrics such as centrality measures, clustering coefficients, and degree distributions.
• To analyze the Integration of Machine Learning Techniques and Knowledge Networks.
• To understand the Knowledge Network Dynamics for collaboration Environments.

Course Outcomes

CO1: Understand the fundamental concepts of knowledge networks.

CO2: Gain practical skills in constructing and visualizing knowledge networks using appropriate tools and techniques.

CO3: Develop ability to perform statistical analyses of knowledge networks using recent graph learning algorithms to identify patterns and insights

CO4: Understand and apply knowledge network principles to related real world problems and applications.

CO5: Evaluate the ethical, policy, and social implications of deploying knowledge networks, ensuring responsible and fair use of knowledge networks.

CO-PO Mapping

Evaluation Pattern: 60/40

Note: Continuous assessments can include quizzes, tutorials, lab assessments, case study and project reviews. Midterm and End semester exams can be a theory exam or lab integrated exam for two hours

1. Network Science by Albert-László Barabási, 1st Edition, 2016.
2. Mining the Social Web: Data Mining Facebook, Twitter, LinkedIn, Instagram, GitHub, and
3. More by Matthew A. Russell, 3rd Edition, 2018.
4. Graph Representation Learning by William L. Hamilton, 1st Edition, 2020.
5. Semantic Web for the Working Ontologist: Effective Modelling in RDFS and OWL by Dean Allemang and James Hendler, 2nd Edition, 2011.
6. Community Detection and Mining in Social Media by Lei Tang and Huan Liu, 1st Edition, 2010.
7. Networks, Crowds, and Markets: Reasoning About a Highly Connected World by David Easley and Jon Kleinberg, 1st Edition, 2010.