Schroedinger's Threshold: When the AUC doesn't predict Accuracy
This addresses a critical issue for researchers and practitioners using AUC in benchmarking, though it is incremental as it focuses on calibration improvements.
The paper tackles the problem that AUC can misrepresent actual accuracy in model evaluation, particularly for faithfulness prediction in generated text, showing it causes significant changes in benchmark rankings. They explore calibration modes to better align with downstream performance.
The Area Under Curve measure (AUC) seems apt to evaluate and compare diverse models, possibly without calibration. An important example of AUC application is the evaluation and benchmarking of models that predict faithfulness of generated text. But we show that the AUC yields an academic and optimistic notion of accuracy that can misalign with the actual accuracy observed in application, yielding significant changes in benchmark rankings. To paint a more realistic picture of downstream model performance (and prepare a model for actual application), we explore different calibration modes, testing calibration data and method.