Position: All Current Generative Fidelity and Diversity Metrics are Flawed
This is a foundational critique for the generative modeling community, highlighting a critical bottleneck in evaluating and applying synthetic data.
The paper argues that all current metrics for evaluating generative model fidelity and diversity are flawed, based on a proposed list of desiderata and sanity checks that reveal issues like lack of robustness and unclear bounds, which hinders practical synthetic data use.
Any method's development and practical application is limited by our ability to measure its reliability. The popularity of generative modeling emphasizes the importance of good synthetic data metrics. Unfortunately, previous works have found many failure cases in current metrics, for example lack of outlier robustness and unclear lower and upper bounds. We propose a list of desiderata for synthetic data metrics, and a suite of sanity checks: carefully chosen simple experiments that aim to detect specific and known generative modeling failure modes. Based on these desiderata and the results of our checks, we arrive at our position: all current generative fidelity and diversity metrics are flawed. This significantly hinders practical use of synthetic data. Our aim is to convince the research community to spend more effort in developing metrics, instead of models. Additionally, through analyzing how current metrics fail, we provide practitioners with guidelines on how these metrics should (not) be used.