Delaying projections in EigenPro-style kernel training cuts the amortized per-batch cost from quadratic to linear in model size, yielding up to hundreds of times faster training at comparable accuracy.
Product kernel interpolation for scalable gaussian processes
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
stat.ML 1years
2024 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Fast training of large kernel models with delayed projections
Delaying projections in EigenPro-style kernel training cuts the amortized per-batch cost from quadratic to linear in model size, yielding up to hundreds of times faster training at comparable accuracy.