Instructor: | A. Erdem Sariyuce (erdem AT buffalo.edu) |

Class hours: | Wed 5:00-7:00, Norton 213 |

Office hours: | Wed 3:30-4:30, Davis 323 |

Graphs are everywhere. Their scale, rate of change, and the irregular nature pose many new challenges. Deep learning has been shown to be successful in a number of domains, ranging from images to natural language processing. However, applying deep learning to the ubiquitous graph data is non-trivial because of the unique characteristics. This seminar course covers recent papers in the last few years about deep learning on graphs. We will consider graph embeddings, knowledge graphs, graph kernels, graph neural networks, graph convolutional networks, graph adversarial methods. Students will learn the literature on deep learning on graphs, understand the state-of-the-art algorithms on various problems, and be familiar with the recent trends.

It is assumed that students have a solid background on discrete mathematics and algorithms. Basic research skills like paper reading, critical thinking, problem solving, report writing, communication, and presentation are important as well.

- Serving as an "explainer" (one time): 35 pts
- Serving as a "listener" (11 times) : 11*5 = 55 pts
- Attending dept talks : 10 pts (more details will be provided)

The final grade is S/U and 80 pts score is needed for an S.

Each week there will be 3 or 4 "explainers". Those students are responsible for presenting the paper and answering the student questions on Piazza. Instructor will pick the presenter among those explainers at the beginning of the class. All explainers must be ready to present. Assignment of the explainers for each week will be done by instructor.

Each week all the students are supposed to read the paper of the week. Each student (except explainers) will ask **a unique question (one question, not less, not more)** on Piazza. Those questions will be answered by the explainers on Piazza and some questions will be selected by them for discussion in class. Questions should be open-ended and provide ground for class discussions, i.e., 'can you explain alg 1?' is not that kind of question. Questions should be posted to Piazza by Monday night, 11.59 pm EST.

The schedule for the first two weeks is below. Rest will be posted soon.

- Feb 1:
**No Class, Instructor is sick**

- Feb 8:

**Course overview by instructor [pdf]**

Machine Learning on Graphs: A Model and Comprehensive Taxonomy Journal of Machine Learning Research 23 (2022) 1-64*by Jason*

- Feb 15:

**Laplacian Eigenmaps for Dimensionality Reduction and Data Representation Neural Computation 2003***by Chandra, Mahima, Ravindra*

- Feb 22:

**DeepWalk: Online Learning of Social Representations SIGKDD 2014***by Hemanth, Shraddha, Yochana*

- Mar 1:

**Revisiting Semi-Supervised Learning with Graph Embeddings ICML 2016***by Abhay, Sravan, Syfullah*

- Mar 8:

**Network Embedding as Matrix Factorization: Unifying DeepWalk, LINE, PTE, and node2vec WSDM 2018***by Abhigna, Kavya, Phaneeswar*

- Mar 15:

**Poincaré Embeddings for Learning Hierarchical Representations NIPS 2017***by Cole, Naveen, Yash*

- Mar 29:

**Deep Graph Infomax ICLR 2019***by Jeremy, Sreeja, Vishal*

- Apr 5:

**Semi-supervised Classification with Graph Convolutional Networks ICLR 2017***by Charan, Lakshay, Saiful, Sujith*

- Apr 12:

**Inductive Representation Learning on Large Graphs NIPS 2017***by Dhanush, Kirthana, Sanjana, Shylesh*

- Apr 19:

**Graph Attention Networks ICLR 2018***by Jason, Sai, Vamsi*

- Apr 26:

**Weisfeiler and Leman Go Neural: Higher-order Graph Neural Networks AAAI 2019***by Anvitha, Mallikharjuna, Pavana, Smriti*

- May 3:

**Combining Label Propagation and Simple Models Out-performs Graph Neural Networks ICLR 2021***by Karthik, Nikhitha, Samar*

- May 10:

**Beyond Homophily in Graph Neural Networks: Current Limitations and Effective Designs NeurIPS 2020***by Niharika, Raakhal, Rounak*