Derives adaptive generalization bounds {c_m / N^{1/(2∨m)}} for digital ML models via new concentration of measure results on finite metric spaces, with c_m = O(sqrt(m)).
Empirical measures: regularity is a counter-curse to dimensionality
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
verdicts
UNVERDICTED 2representative citing papers
Scale-function characterizations of pre-infimum paths under two decompositions for spectrally negative Lévy processes give Doob h-transform laws and explicit drawdown distributions on components.
citing papers explorer
-
Tighter Learning Guarantees on Digital Computers via Concentration of Measure on Finite Spaces
Derives adaptive generalization bounds {c_m / N^{1/(2∨m)}} for digital ML models via new concentration of measure results on finite metric spaces, with c_m = O(sqrt(m)).
-
Conditional Path Decomposition at the Infimum and Maximum Drawdowns for Spectrally Negative L\'{e}vy Processes
Scale-function characterizations of pre-infimum paths under two decompositions for spectrally negative Lévy processes give Doob h-transform laws and explicit drawdown distributions on components.