This is loserlulin's summary about papers on Byzantine attack and defense.
Some equations may not be displayed properly in github. You can download TYPORA or just browse the PDF file we supply.
1. Machine Learning with Adversaries: Byzantine Tolerant Gradient Descent (NeurIPS 2017)
2. Byzantine-Robust Distributed Learning: Towards Optimal Statistical Rates (ICML 2018)