Max Jerdee
PhD Student, University of Michigan
Wed, Oct 11, 2023
2:00 PM UTC
Wed, Oct 11, 2023
2:00 PM UTC
In-person
4 Thomas More St
London E1W 1YW, UK
London E1W 1YW, UK
The Roux Institute
Room
100 Fore Street
Portland, ME 04101
Portland, ME 04101
Network Science Institute
2nd floor
2nd floor
Network Science Institute
11th floor
11th floor
177 Huntington Ave
Boston, MA 02115
Boston, MA 02115
Room
58 St Katharine's Way
London E1W 1LP, UK
London E1W 1LP, UK
Talk recording
We discuss simple yet impactful adjustments to two popular tools of network science. In community detection, Normalized Mutual Information is often used as a similarity measure to assess the quality of an inferred community structure against "ground truth" communities. We describe two drawbacks of this measure: bias towards excessive groups and bias in normalization and propose remedies to both. In ranking, the Bradley-Terry model (commonly known as the basis for Elo scores in chess) is widely applied to infer rankings in many contexts including sports and animal and social interactions. We propose a generalization of this model which can infer the role of luck and the depth of competition in these various hierarchies.
About the speaker
Max is a 4th year Physics PhD student at the University of Michigan in Mark Newman’s group. He graduated with a BA in Physics from Princeton in 2020, and his research interests since have wandered from astrophysics to high energy theory, but he now mostly works on using physical ideas to explore information theory and rankings on networks.
Share this page: