These tools and metrics are designed to help AI actors develop and use trustworthy AI systems and applications that respect human rights and are fair, transparent, explainable, robust, secure and safe.
Inferred Average Precision (infAP)
Inferred Average Precision (infAP) is a statistical measure used to estimate the average precision of a retrieval system when relevance judgments are incomplete. It addresses the issue of having a limited amount of labeled data by "inferring" the full collection's average precision from a subset of documents. Essentially, infAP uses a sampling method to estimate the average precision from a smaller, labeled dataset.
infAP can indirectly reinforce Robustness by giving a reliable, sample-efficient signal of retrieval quality even when full relevance labels are missing, so sudden drops in inferred average precision flag distribution shifts before they degrade user experience.
About the metric
You can click on the links to see the associated metrics
Objective(s):
Purpose(s):
Lifecycle stage(s):
Target users:
