TurboTest decouples ML throughput prediction from termination classification to enable early stopping of speed tests, delivering 1.8-4.4x higher data savings than BBR baselines on 1 million M-Lab tests while lowering median error.
Extracting diverse-shapelets for early classification on time series.World Wide Web, 23(6):3055–3081
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.NI 1years
2025 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
TURBOTEST: Learning When Less is Enough through Early Termination of Internet Speed Tests
TurboTest decouples ML throughput prediction from termination classification to enable early stopping of speed tests, delivering 1.8-4.4x higher data savings than BBR baselines on 1 million M-Lab tests while lowering median error.