Open source implementations and research software
Below are implementations of algorithms and methods developed in our research. All code is available on GitHub.
A fast and scalable kernel solver for large-scale machine learning problems. EigenPro efficiently handles kernel methods through eigendecomposition-based approximations, enabling training on datasets with millions of samples.
View on GitHubImplementation of Recursive Feature Machines (RFM), a novel approach to feature learning that recursively constructs features through iterative optimization. This method provides an alternative to traditional neural network architectures.
View on GitHub