overparameterization

[ICML 2021] "Do We Actually Need Dense Over-Parameterization? In-Time Over-Parameterization in Sparse Training" by Shiwei Liu, Lu Yin, Decebal Constantin Mocanu, Mykola Pechenizkiy

This project outlines 4 experiments to explore the effects of several settings on the bias-variance tradeoff curve

An intuitive derivation of smoothing splines from variational calculus, demonstrating their relationship to reproducing kernel Hilbert spaces (RKHS) and regularized neural networks.

Code and figures for To Grok Grokking: Provable Grokking in Ridge Regression (ICML 2026), with ridge-regression, random-feature, NTK-style, and fully trained ReLU experiments.

A reproduction of the results from Ribeiro & Schön (2023) on adversarial attacks in overparameterized linear regression. Developed as an undergraduate research project funded by the PICME-CNPq Scientific Initiation program.