Last released Sep 28, 2026
Open-source detection of data poisoning and backdoor attacks in ML classification pipelines, validated on healthcare imaging and general-purpose benchmarks.