SCOOTER supplies best-practice guidelines, open tools, and a 3K-image benchmark with 34K+ human ratings showing that six tested unrestricted attacks produce images humans can detect as fake.
Title resolution pending
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
representative citing papers
Derives exact marginal likelihood under finite-support Huber contamination via Dirichlet-Beta priors and dynamic programming over count allocations.
citing papers explorer
-
SCOOTER: A Human Evaluation Framework for Unrestricted Adversarial Examples
SCOOTER supplies best-practice guidelines, open tools, and a 3K-image benchmark with 34K+ human ratings showing that six tested unrestricted attacks produce images humans can detect as fake.
-
Marginal likelihoods for finite-support Huber contamination
Derives exact marginal likelihood under finite-support Huber contamination via Dirichlet-Beta priors and dynamic programming over count allocations.