As a form of unsupervised learning, clustering is an inherently ill-defined problem. Often times a researcher may employ spectral clustering to find the “true” communities in a network, but what if there are two different, yet correct, “truths”?
This paper was written for a final project of CS590OP: Numerical Optimization at UMass Amherst. It contains an overview of simulated annealing and attempts to recover Stochastic Block Model parameters via the method.
Materials used in the Fall 2023/2024 MS Boot Camp at Duke University. Code used to create this material can be found here.
If a player performs perfectly in a matching game, how many turns will they need to win?
This tiny survey paper was written as a culmination of readings and conversions from my independent study on statistics and networks with Vince Lyzinski.