1 paper
Emre Kavak, Tom Nuno Wolf, Christian Wachinger
Dataset bias often leads deep learning models to exploit spurious correlations instead of task-relevant signals. We introduce the Standard Anti-Causal Model (SAM), a unifying causa…