Conformal Prediction in Learning Under Privileged Information Paradigm with Applications in Drug Discovery
This work addresses uncertainty quantification in drug discovery and other domains using privileged information, but it is incremental as it combines existing methods.
The paper applied conformal prediction with SVM+ in the learning under privileged information paradigm to MNIST and drug discovery datasets, showing valid models and improved efficiency over standard SVM, though gains in drug discovery were limited, while enabling valid prediction intervals at specified significance levels.
This paper explores conformal prediction in the learning under privileged information (LUPI) paradigm. We use the SVM+ realization of LUPI in an inductive conformal predictor, and apply it to the MNIST benchmark dataset and three datasets in drug discovery. The results show that using privileged information produces valid models and improves efficiency compared to standard SVM, however the improvement varies between the tested datasets and is not substantial in the drug discovery applications. More importantly, using SVM+ in a conformal prediction framework enables valid prediction intervals at specified significance levels.