Single-round probabilistic bipartite matching algorithms with degree-biased sampling and random thinning achieve higher expected matching sizes and extend stability regions under communication constraints in data center networks.
Rail-only: A low-cost high-performance network for training LLMs with trillion parameters
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.DC 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Bipartite matching under communication constraints
Single-round probabilistic bipartite matching algorithms with degree-biased sampling and random thinning achieve higher expected matching sizes and extend stability regions under communication constraints in data center networks.