Shan Leng

I am currently a research assistant working under the supervision of Dr. Edward Ott and Dr. Brian Hunt.

I am currently studying the neural scaling problems related to ridge regression-based neural scaling theory for dynamical time-series forecasting.

• Developing scaling theory of generalization error with respect to training length, number of trainable parameter, regularization strength, noise strength, and time-correlation in noise.
• Conduct numerical experiment with Reservoir computing, a type of single-layer recurrent neural network to verify the theory

I learned about a type of graph called Adinkra, used to explore the properties of a supersymmetric model in summer of 2023.

I am currently finishing a paper coauthored by Dr. Jim Gates about using Algebraic Geometry-based method to search for new relevant interactions in supersymmetric theory through adinkra morphism.

• Given a talk titled "Computational Algebraic Geometrical Methods for The Supersymmetry Off-Shell Problem" at an informal seminar series hosted by Dr. Jim Gates

• Joined an interdisciplinary research group studying the effect of focus of public communication on vaccine hesitancy spread using an Ising Model network-based approach

• I taught the lab and recitation sections of Physics 131 and 132, the physics for biologist.