Introduces staleness-bucket aggregation with padding for distributed perceptron and proves finite-horizon mistake bounds where delay affects only mean staleness and noise adds a sqrt(horizon) term.
Online learning under delayed feedback,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Distributed Perceptron under Bounded Staleness, Partial Participation, and Noisy Communication
Introduces staleness-bucket aggregation with padding for distributed perceptron and proves finite-horizon mistake bounds where delay affects only mean staleness and noise adds a sqrt(horizon) term.