Interpretability BasedOn Using Learnable If Statements for Interpretability sparsity for interpretability Leveraging sample sparsity to improve interpretability of neural networks Robustness Data Leakage of LoRA in federated training This article highlights the potential risks of reconstructing private data from the gradients shared in Federated Learning, especially when using the LoRA finetuning technique. Certified Robust Neural Network Certify Robustness using median neurons
