Implementation of a Physics Informed Neural Network (PINN) written in Tensorflow v2, which is capable of solving Partial Differential Equations.

In this project, I implement an enhanced active contour method that uses discrete wavelet transform for energy minimization to increase the accuracy.

This repository contains wrapper scripts for running transition state and IRC (Intrinsic Reaction Coordinate) calculations using Sella and IRC ASE optimizers for the Sella package.

Official implementation of "Unraveling the Hessian: A Key to Smooth Convergence in Loss Function Landscapes"

The implementation of advanced mathematical optimization methods

A curated repository of papers on the Hessian of deep neural networks.

This is my project about liver vessel segmentation in CT images based on hessian matrix and U-Net networks

Dual number classes for automatic differentiation including for higher order derivatives

Learning Network using Hessian Optimization in PyTorch

Code related to the paper: Hartoyo, A., Argasiński, J., Trenk, A., Przybylska, K., Błasiak, A., & Crimi, A. (2025). Synergistic eigenanalysis of covariance and Hessian matrices for enhanced binary classification on health datasets. Computers in Biology and Medicine, 190, 109985.

PyTorch tools for Hessian-related operations

A JAX + SymPy based project that explores optimization in ML using hessian matrix and graidents.