DRSGD-ByMI identifies Byzantine machines via sample-splitting score statistics with FDR control, then prunes them to recover sufficient connectivity and achieve order-optimal convergence rates identical to standard decentralized SGD.
Agnostic estimation of mean and covariance
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
verdicts
UNVERDICTED 2representative citing papers
A novel robust asynchronous Q-learning algorithm achieves finite-time convergence rates that match clean-data bounds up to an additive term proportional to the corruption fraction, with a matching information-theoretic lower bound.
citing papers explorer
-
Toward Exact Convergence in Byzantine-Robust Decentralized Learning: A Statistical Identification Approach
DRSGD-ByMI identifies Byzantine machines via sample-splitting score statistics with FDR control, then prunes them to recover sufficient connectivity and achieve order-optimal convergence rates identical to standard decentralized SGD.
-
Corruption-Tolerant Asynchronous Q-Learning with Near-Optimal Rates
A novel robust asynchronous Q-learning algorithm achieves finite-time convergence rates that match clean-data bounds up to an additive term proportional to the corruption fraction, with a matching information-theoretic lower bound.