# Topic proposals and voting

Please add to this list ideas and proposals for topics to do. People can submit their preferences using the typical emoji reactions for project allocation (in REG):
* Really want to have this covered: use a `:smile:`/`:smile_cat:` reaction: :smile: / :smile_cat:
- Happy to learn more: use a `:+1:` reaction: :+1:
- Not interested: use a `:-1` reaction: :-1:

Note that proposed topics can be big or small. For larger topics, they may span over several sessions.

**You don't have to be an expert in the topic to present!** The only requirement is that you go and digest some material, and re-present that back to the group. (You can even include your questions!)

If you'd like to present a topic that interests you below, please message Nathan or Oliver in REG.

## Potential speakers (feel free to add your name here!)

- Nathan
- Levan
- Oliver
- Edmund (happy to speak about maths anytime)
- Boyko
- Praveen

## Candidates for next resource

- The [Graph Representation Learning Book](https://www.cs.mcgill.ca/~wlh/grl_book/) by Will Hamilton [VOTE HERE: ]
- Renowned as one of the better textbooks in the field
- We can pick and choose topics based on our interests (start with a re-visit to the fundamentals)
- Various topics in the [Geometric Deep Learning lectures](https://geometricdeeplearning.com/lectures/)
- We didn't cover any of the side-topics (Groups/Grids/Gauges/High-dimensional learning), as they get pretty maths-heavy!
- Some cool-looking practicals in here on GNNs!
- Going back through the Phi-ML Hands on sessions on GNNs
- CS224W: Stanford GNN course

## Application sessions

- Boyko: Temporal graphs for traffic + unsupervised learning
- Praveen talks us through his notebook!

## Topics

### Graph theory

- [ ] Linear models
- [ ] HyperGraphs
- [x] Are Transformers GNNs?
- [ ] What are semantic knowledge networks?
- [ ] Are CNNs a special case of GNNs?

### Geometric deep learning lectures
- [ ] Lecture 2: High-Dimensional Learning
- [recording of original](https://youtu.be/plIJYzVKfdI)
- [ ] Lecture 3: Geometric Priors I
- [recording of original](https://youtu.be/qEjWMhRlXgY)
- [ ] Lecture 4: Geometric Priors II
- [recording of original](https://youtu.be/DpnA8NNUtyU)
- [x] Lecture 5: Graphs & Sets I :+1: [Nathan]
- [recording of original](https://youtu.be/J2bLt3-SSpg)
- [x] Lecture 6: Graphs & Sets II :+1:
- [recording of original](https://youtu.be/HvQw7Zq1jtU)

... more on https://geometricdeeplearning.com/lectures/

### Phi-ML Hands on sessions on GNNs

- Playlist: www.youtube.com/@mlds_seminar
- [Lecture 1](https://youtu.be/D_QZy_s5HRI?si=lPdebnQ-eEzjNb7k)
- [Lecture 2](https://youtu.be/XqtZrIQwsa8?si=1araIMY4hRs8Eqja)
- [Lecture 3](https://youtu.be/Vzrnr8UDY4s?si=lch-4j-zpqdmi5-X)
- [Lecture 4](https://youtu.be/mh_oCB9O4rA?si=yuI5NnrzKws_9-_i)

### Applications

- [ ] Learning Mesh-based Simulations with Graph Neural Networks
- [website](https://sites.google.com/view/meshgraphnets)
- [ ] Clustering/Community detection using GNNs :smile_cat:

### Specific papers:

- [ ] [GraphCast: Learning skillful medium-range global weather forecasting
](https://www.science.org/doi/10.1126/science.adi2336)
- [ ] [How Powerful are Graph Neural Networks?](https://arxiv.org/abs/1810.00826)

## Material and links

- [Syllabus for GNN course on Udemy](https://www.udemy.com/course/graph-neural-network/?utm_source=adwords&utm_medium=udemyads&utm_campaign=Webindex_Catchall_la.EN_cc.UK&utm_term=_._ag_114213220700_._ad_532713168385_._kw__._de_c_._dm__._pl__._ti_dsa-392284169515_._li_9045997_._pd__._&matchtype=&gad_source=1&gclid=CjwKCAjwnOipBhBQEiwACyGLug-xeOVz7tqOCRBlit0f6pCfNpIptrwELo0YkCfsfZhzjrv3_TnS6RoC0XsQAvD_BwE)
- [Syllabus for GNN course from UPenn](https://gnn.seas.upenn.edu/lectures/)
- [DeepFindr Short GNN video series](https://www.youtube.com/watch?v=fOctJB4kVlM&list=PLV8yxwGOxvvoNkzPfCx2i8an--Tkt7O8Z&index=1&ab_channel=DeepFindr)