FedVal: Different good or different bad in federated learning
Read more at
https://arxiv.org/abs/2306.04040
Thesis project investigating security against attacks and failures in federated learning, proposing a new innovative approach to create a safe and balanced model while training in a decentralised setting. Thesis got accepted to the USENIX security symposium 2023 which is one of the top 10 computer science conferences. See link above for the paper.
Data, Technology
Research
IT & Software, Innovation, Research & Development
Better Quality
Practioner
Prediction, Vision
CNN, DNN, Machine Learning