Param Mody

collaborators

3 papers

cs.LG2026

A Minimal ReLU Classifier on Non-Linearly Separable Data: Convergence and Implicit Bias at and beyond the Edge of Stability

Quan Nguyen, Robin Yadav, Param Mody +1

Gradient descent (GD) with large learning rates can successfully optimize neural networks even though classical smoothness-based learning rate conditions are violated. Yet theory f…

math.AG2026

Expressivity of Shallow Neural Networks Over Finite Fields

Maksym Zubkov, Carol Wu, Shiwei Yang +2

We study the expressivity of shallow polynomial neural networks (PNNs) with monomial activation functions over finite fields. For a given architecture, we define a neuromanifold as…

math.CO2026

Improved bounds for lines and $1$ -separated sets in Euclidean Ramsey theory

Gabriel Currier, Param Mody, Zehan Xie +1

Let $K$ be a $1$ -separated set of diameter at most $R - 1$ , and let $ℓ_{m}$ denote a collection of $m$ points on a line, with consecutive points of distance $1$ apart. Conlon and Fo…

A Minimal ReLU Classifier on Non-Linearly Separable Data: Convergence and Implicit Bias at and beyond the Edge of Stability

Expressivity of Shallow Neural Networks Over Finite Fields

Improved bounds for lines and 1-separated sets in Euclidean Ramsey theory

Improved bounds for lines and $1$ -separated sets in Euclidean Ramsey theory