CS F426: Graph Mining (Fall 2023,BITS Pilani)
Undergraduate/Post-Graduate Course, BITS Pilani, CSIS, 2023
Course designer: Vinti Agarwal
Time: TThS 11-11:50am
Course material can be found here
Do You Know?
- Google Maps uses Graphs to find the shortest route between point A and B.
- Youtube recommends videos based on knowledge graphs
- Facebook “People you may know” uses Social Graph for friend suggestions
- Amazon recommendations based on product reviews graph
- And many more
To know more, I recommend to browse:
- https://gm-neurips-2020.github.io/
- https://research.google/teams/graph-mining/
- https://developers.facebook.com/docs/graph-api/
- https://economicgraph.linkedin.com/
- https://ai-med.io/literature/graph-based-deep-learning-for-medical-diagnosis-and-analysis/
CS426@Pilani (First time introduced in Fall 2021)
Graph mining CSF426 was introduced for the first time at Pilani campus in the fall 2021 semester and has been taught twice till May 2023. It was liked by previous batches primarily not because it has tremendous applications in the data science domain, but due to the learning style with examples and coding exercise. The fun part of the course is every two lectures followed by a small programming exercise, sometimes with missing lines of code type fun quiz. Coding enthusiasts can take this opportunity to test their skills and earn scores in the easiest way.
Unique features of the course
- Open class discussion once in a month with faculty and peers.
- Peers feedback on Quiz or assignment
- Learning with examples and coding
Who should take this course?
Anyone who wishes to learn:
- state-of-the-art graph based machine learning algorithms and deep learning algorithms.
- Python implementation of basic graph algorithms (e.g. pagerank, spectral clustering, laplacian etc) from scratch
- Handling of unstructured datasets originated from multiple sources and their transformation into graphs (Case studies include social and medical data)
- Working details of graph neural network architectures and its application in prediction and classification tasks