2 papers
cs.IT2026
Communication-Efficient Distributed Training via Ring-Based Coded Approximate All-Reduce
Sifat Munim, Aditya Ramamoorthy
Ring All-Reduce is a widely used protocol for aggregating gradients across workers in large-scale distributed training. For a system with workers, its normalized per-worker com…
cs.IT2026
Communication-Efficient Approximate Gradient Coding
Sifat Munim, Aditya Ramamoorthy
Large-scale distributed learning aims at minimizing a loss function that depends on a training dataset with respect to a -length parameter vector. The distributed cluster ty…