Do We Really Need to Design New Byzantine-robust Aggregation Rules?
Minghong Fang
Network and Distributed System Security (NDSS) Symposium 2025 · Day 3 · Federated Learning 2
This talk, presented by Minghong Fang at the NDSS Symposium, challenges the prevailing trend in Federated Learning (FL) security research: the continuous development of increasingly complex Byzantine-robust aggregation rules. Fang argues that instead of constantly designing novel defenses, the community should focus on enhancing the robustness of existing, foundational aggregation methods. The core premise is that by strategically manipulating the input distribution to these foundational rules, their efficacy against advanced poisoning attacks can be significantly improved, obviating the need for more intricate and often vulnerable new designs.
AI review
Solid academic ML security research that makes a clear, testable argument — stop chasing complexity, fix the fundamentals — and backs it with reasonable empirical breadth (12 attacks, 10 baselines, 6 datasets). Not groundbreaking, but honest and well-scoped work that belongs in an academic venue.