Benchmarking non-conformity score functions in conformal prediction
arXiv:2605.24983v1 Announce Type: new Abstract: Conformal prediction is a useful and versatile alternative to model calibration in machine learning classification. It replaces single-class prediction