FFDCI trains a feature-based quantile error predictor and then applies an online adaptive offset, claiming asymptotic coverage and shorter intervals than existing conformal baselines on 12 datasets.
The authors argue that this indicator function is excessively non-smooth and propose replacing it with a sigmoid function
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
Feature Fitted Online Conformal Prediction for Deep Time Series Forecasting Model
FFDCI trains a feature-based quantile error predictor and then applies an online adaptive offset, claiming asymptotic coverage and shorter intervals than existing conformal baselines on 12 datasets.