robust-aggregation

Sovereign Map is a production-grade, Byzantine-tolerant Federated Learning framework. Utilizing the Mohawk Protocol for streaming aggregation, it achieves a 224x memory reduction, enabling secure orchestration of 100M+ nodes via TPM 2.0 hardware-rooted trust. Features full-stack observability with Prometheus & Grafana, built-in tokenomics telemetry

Federated Fusion CUDA primitives for FL server-side aggregation, optimization, compression, and robustness.

Reproducible TDAW research lab for history-aware Byzantine-robust federated MNIST training, with a Rust core, Flower modes, and a Next.js dashboard.

Train one machine-learning model across many parties who don't trust each other. A dishonest minority can't skew the result, and every party computes byte-for-byte identical output on any CPU or architecture (big-endian included) — so each round ships a receipt anyone can re-check offline and reproduce exactly. Rust; drop-in for Flower.

Offline ensemble-combiner benchmark: competence weighting and robust aggregation recover a known Bayes signal where naive averaging collapses -- a 2x2 dissociation (weighting x aggregation) plus a diversity control, numpy-only, no downloads.

Empirical study of the Privacy–Robustness–Performance trilemma in Federated Learning: combining DP-SGD, FLTrust Byzantine-robust aggregation, and Top-k compression across 8 configurations on MNIST, simulated with Flower.

byzantine-tolerant aggregation primitives for federated learning, with deterministic and reputation-aware options.