Get Started
Home
Topics
Search
Library
Research questionHow can medical-image counterfactuals reveal a classifier’s decisions without inheriting a generator’s biases?Counterfactual images can show what changes would flip a diagnosis, but an auxiliary generator may impose its own inductive biases. This makes it difficult to determine whether an apparent explanation reflects the classifier’s evidence or the generator’s image-synthesis behavior.
AI
Computer Vision
Evaluation & Benchmarks
Health
Image & Video Processing
Machine Learning
Latest papersRecent research connected to this question, newest first.Generating Medical Image Counterfactuals using Causal ExplanationsThe source concerns black-box medical image classifiers and counterfactual images for auditing their decisions. It reports evidence from real-world medical imaging datasets for a deterministic, no-additional-training approach with user-specified regions of interest, including comparisons with generative baselines.research paper · Sep 2, 2026
Related questions
How can we generate realistic, minimally edited visual counterfactuals without inheriting classifier shortcuts?How can counterfactual image edits be audited for a shared latent explanation across regions?How can we reliably detect memorized patient images in medical generative models at scale?How can we compare medical VQA models when generation obscures answer signal and biases answer positions?