Introduces SOBBO, plus an estimate-then-differentiate surrogate method and a conservative-field gradient interpolation method, and shows the latter is more accurate and robust in scarce-data experiments.
The true gradient of θ is estimated by the SimOpt testbed
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2024 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Offline Stochastic Optimization of Black-Box Objective Functions
Introduces SOBBO, plus an estimate-then-differentiate surrogate method and a conservative-field gradient interpolation method, and shows the latter is more accurate and robust in scarce-data experiments.