lipschitz

Build and train Lipschitz constrained networks: TensorFlow implementation of k-Lipschitz layers

Constrained deep learning is an advanced approach to training deep neural networks by incorporating domain-specific constraints into the learning process.

Code for paper "ECLipsE: Efficient Compositional Lipschitz Constant Estimation for Deep Neural Networks"

Quantum faireness verifying algorithm based on calculating Lipschitz constant, an implementaion of paper: https://arxiv.org/pdf/2207.11173.pdf💿

Reference implementation for fixed-threshold fairness auditing and ReLiF multi-task learning experiments