Stochastic Process Library for Python

Neural Tangent Kernel (NTK) module for the scikit-learn library

A benchmarking library for quantum and classical machine learning, with specialized support for evaluating kernel methods.

This project simulates an experiment using Adaptive Random Fourier Features Kernel Least Mean Squares (ARFF-KLMS) to approximate non-linear functions. The algorithm adapts kernel bandwidth online, enhancing tracking and convergence in non-stationary environments. The experiment calculates and plots MSE over iterations for performance evaluation.

This project demonstrates a simulation of an experiment using Random Fourier Features (RFF) to approximate a non-linear function. The experiment is designed to be repeated multiple times, and the mean squared error (MSE) is calculated and plotted over iterations to evaluate the performance of the model.

Nonlinear model reduction for operator learning

This resp presents a probabilistic and online forecasting model. In detail, a deep kernel is proposed by integrating the deep soft Spiking Neural Networks into the Gaussian kernel, which is then applied to perform sparse Gaussian Process regression.

Classic methods on digit recognition. As part of the MITx course on machine learning with Python - from linear models to deep learning

Official repository for the paper "Time Series Kernels based on Nonlinear Vector AutoRegressive Delay Embeddings" (NeurIPS 2023)

Large-scale, multi-GPU capable, kernel solver

Causal Modeling with Stationary Diffusions, AISTATS 2024

Exploring "variability collapse" in shallow neural networks

Code for reimplementing "Stein Variational Gradient Descent (SVGD)" method.

dckernel package implementing dcMMD and dcHSIC from Robust Kernel Hypothesis Testing under Data Corruption, by Schrab and Kim

KSDAgg package implementing the KSDAgg test proposed in KSD Aggregated Goodness-of-fit Test by Schrab, Guedj and Gretton: https://arxiv.org/abs/2202.00824 NeurIPS 2022

AggInc package implementing the MMDAggInc, HSICAggInc and KSDAggInc tests proposed in Efficient Aggregated Kernel Tests using Incomplete U-statistics by Schrab, Kim, Guedj and Gretton: https://arxiv.org/abs/2206.09194 NeurIPS 2022

MMD-FUSE package implementing the MMD-FUSE test proposed in MMD-FUSE: Learning and Combining Kernels for Two-Sample Testing Without Data Splitting by Biggs, Schrab, and Gretton: https://arxiv.org/abs/2306.08777

MMDAgg package implementing the MMDAgg test proposed in MMD Aggregated Two-Sample Test by Schrab, Kim, Albert, Laurent, Guedj and Gretton: https://arxiv.org/abs/2110.15073

Code for reproducible experiments presented in KSD Aggregated Goodness-of-fit Test. https://arxiv.org/abs/2202.00824 NeurIPS 2022

dpkernel package implementing dpMMD and dpHSIC from Differentially Private Permutation Tests: Applications to Kernel Methods, by Kim and Schrab