GIFT computes a gradient-informed direction based on how the training loss gradient changes with the assumed noise level and line-searches along it in situ, improving accuracy under noise misspecification.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.NE 1years
2025 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
In situ fine-tuning of in silico trained Optical Neural Networks
GIFT computes a gradient-informed direction based on how the training loss gradient changes with the assumed noise level and line-searches along it in situ, improving accuracy under noise misspecification.