Student Lecture

Author

Sadamori Kojaku

Published

August 25, 2026

This semester you will give a 20-minute lesson once. The lesson should be on something that we do not cover in this course.

Rules

  • Once, for 15-20 minutes.
  • Individually, or as a team of up to three people.
  • This accounts for 10 points of the final grade, and the same score goes to all members of the team.
  • The class has to do something while you’re up there, and a lecture alone doesn’t pass.
  • You select one topic from the pools below and one slot from the table, with each team taking exactly one topic and one slot.
  • Fill out the sign-up sheet with your name, topic, and slot. The first team to write on each line claims the slot.
  • If you would like any feedback on your plan, please feel free to email me. If needed, we can also set up a time to meet during office hours or at another time to discuss it. I will stop reading plans one week before your date.
  • You can bring in your own topic instead, making sure that the topic is not already covered anywhere in the course. Tell me what the class will do.

What “do something” means

Good:

  • Everyone fills out the worksheet.
  • Everyone should calculate a small example by hand.
  • Everyone writes down their guesses before seeing the answers.
  • Everyone plays a short game.
  • Write some code live and let the class tell you what to test.

Not enough:

  • Any questions?
  • A poll that changes nothing.
  • A demo for the class to watch.
  • A quiz will be given at the end.

Ask yourself: If your class were at home, would your 20 minutes look the same? If so, the activity is decoration.

Your date

Each team takes one of the 13 slots. Write down the slot you want on the sign-up sheet. Each session runs 20 minutes from the start of a regular class. If you form a team, some slots stay open.

Slot Date Lecture that day
01 Sep 17 (Thu) Friendship Paradox
02 Sep 22 (Tue) Friendship Paradox
03 Sep 29 (Tue) Clustering
04 Oct 1 (Thu) Clustering
05 Oct 6 (Tue) Clustering
06 Oct 20 (Tue) Centrality
07 Oct 22 (Thu) Centrality
08 Oct 29 (Thu) Random Walks
09 Nov 3 (Tue) Random Walks
10 Nov 10 (Tue) Graph Embedding
11 Nov 12 (Thu) Graph Embedding
12 Nov 19 (Thu) Graph Neural Networks
13 Nov 24 (Tue) Graph Neural Networks

Topic pools

There are 26 areas, none of which will be covered elsewhere in the course. Each area is broader than a single lecture, so you choose the angle. The references are a place to start reading.

Kinds of network

  • SL-01 Bipartite networks. There are two types of nodes. What does a projection onto one type do? Start: Newman, Networks.
  • SL-02 Signed networks. Links representing a “friend” or “enemy”, and structural balance. Start: Cartwright and Harary, Psychological Review 1956.
  • SL-03 Higher-order networks. Hypergraphs and simplicial complexes can represent interactions among multiple nodes, where a single interaction can connect more than two nodes. Start: Battiston et al., Physics Reports 2020.
  • SL-04 Multilayer and interdependent networks. Multiple networks can be located on the same nodes, and multiple networks can depend on each other. Start: Kivela et al., Journal of Complex Networks 2014. Buldyrev et al., Nature 2010.

Dynamics on networks

The structure is fixed.

  • SL-05 Epidemic models. The SIR and SIS models on networks, and the epidemic threshold. Start: Pastor-Satorras and Vespignani, PRL 2001.
  • SL-06 Contagion and cascades. Threshold models, complex contagion, and global cascades. Start: Watts, PNAS 2002. Centola and Macy, AJS 2007.
  • SL-07 Opinion dynamics. The voter model and its variants. Start: Castellano, Fortunato and Loreto, Reviews of Modern Physics 2009.
  • SL-08 Synchronization. Coupled oscillators: which structures synchronize easily? Start: Arenas et al., Physics Reports 2008.
  • SL-09 Evolutionary games. Cooperation and defection among players on networks. Start: Nowak and May, Nature 1992.

Dynamics of networks

The structure itself changes.

  • SL-10 Temporal networks. Links that appear and disappear; time-respecting paths. Start: Holme and Saramaki, Physics Reports 2012.
  • SL-11 Network growth. How does a real network grow? What factors determine where new links attach? Start: Papadopoulos, Kitsak, Serrano, Boguna and Krioukov, Nature 2012. Leskovec, Kleinberg and Faloutsos, KDD 2005. Bianconi and Barabasi, EPL 2001.
  • SL-12 Adaptive networks. The structure and the node states change each other. Start: Gross and Blasius, Journal of the Royal Society Interface 2008.

Control and navigation

  • SL-13 Controllability and observability. Which node to drive? Where to place the sensors? Start: Liu, Slotine and Barabasi, Nature 2011 and PNAS 2013.
  • SL-14 Navigability. Finding short paths with only local information. Start: Kleinberg, Nature 2000.

Structure

  • SL-15 Network motifs. Small subgraphs occurring more often than by chance. Start: Milo et al., Science 2002.
  • SL-16 Statistical models of networks. Exponential random graph models. Start: Robins et al., Social Networks 2007.
  • SL-17 Core-periphery structure and node roles. Positioning beyond group membership. Start: Borgatti and Everett, Social Networks 2000. Guimera and Amaral, Nature 2005.
  • SL-18 Limits of community detection. Detectability and the choice of the number of groups. Start: Decelle et al., PRL 2011.

Machine learning on networks

  • SL-19 Link prediction. Predicting missing or unformed links, and evaluating predictions. Start: Liben-Nowell and Kleinberg, JASIST 2007. Zhang and Chen, NeurIPS 2018 (SEAL).
  • SL-20 Knowledge graphs. Networks with typed links, and how to embed them. Start: Bordes et al., NeurIPS 2013.
  • SL-21 Limits of graph neural networks. Over-squashing, expressivity, what node embeddings cannot say about pairs, and explaining predictions. Start: Alon and Yahav, ICLR 2021. Zhang and Chen, NeurIPS 2018 (SEAL). Ying et al., NeurIPS 2019.

Visualization

  • SL-22 Network layout. Force-directed placement and alternatives. Start: Fruchterman and Reingold, Software: Practice and Experience 1991.
  • SL-23 Large networks on a page. Aggregation, bundling and matrix views. Start: Ghoniem, Fekete and Castagliola, InfoVis 2004. Holten, IEEE TVCG 2006.

Where the network comes from

  • SL-24 Inferring networks from data. Correlation network and backbone extraction. Start: Zalesky, Fornito and Bullmore, NeuroImage 2012. Serrano, Boguna and Vespignani, PNAS 2009.
  • SL-25 Network reconstruction. Estimating links when only node totals are observed, like banks and their balance sheets. Start: Squartini, Caldarelli, Cimini, Gabrielli and Garlaschelli, Physics Reports 2018. Cimini et al., Nature Reviews Physics 2019.

Networks and society

  • SL-26 Recommendation, ranking and misinformation. What do link algorithms do to a group? How does misinformation spread? Start: Karimi et al., Scientific Reports 2018. Vosoughi, Roy and Aral, Science 2018.