Timezone: »

How Faithful is your Synthetic Data? Sample-level Metrics for Evaluating and Auditing Generative Models
Ahmed Alaa · Boris van Breugel · Evgeny S. Saveliev · Mihaela van der Schaar

Tue Jul 19 03:30 PM -- 05:30 PM (PDT) @ Hall E #218

Devising domain- and model-agnostic evaluation metrics for generative models is an important and as yet unresolved problem. Most existing metrics, which were tailored solely to the image synthesis setup, exhibit a limited capacity for diagnosing the different modes of failure of generative models across broader application domains. In this paper, we introduce a 3-dimensional evaluation metric, (α-Precision, β-Recall, Authenticity), that characterizes the fidelity, diversity and generalization performance of any generative model in a domain-agnostic fashion. Our metric unifies statistical divergence measures with precision-recall analysis, enabling sample- and distribution-level diagnoses of model fidelity and diversity. We introduce generalization as an additional, independent dimension (to the fidelity-diversity trade-off) that quantifies the extent to which a model copies training data—a crucial performance indicator when modeling sensitive data with requirements on privacy. The three metric components correspond to (interpretable) probabilistic quantities, and are estimated via sample-level binary classification. The sample-level nature of our metric inspires a novel use case which we call model auditing, wherein we judge the quality of individual samples generated by a (black-box) model, discarding low-quality samples and hence improving the overall model performance in a post-hoc manner.

Author Information

Ahmed Alaa (UCLA)
Boris van Breugel (University of Cambridge)
Evgeny S. Saveliev (University of Cambridge)
Mihaela van der Schaar (University of Cambridge and UCLA)

Related Events (a corresponding poster, oral, or spotlight)

More from the Same Authors