Mitigating attribute amplification in counterfactual image generation

arXiv:2403.09422v1 Announce Type: new
Abstract: Causal generative modelling is gaining interest in medical imaging due to its ability to answer interventional and counterfactual queries. Most work focuses on generating counterfactual images that look plausible, using auxiliary classifiers to enforce effectiveness of simulated interventions. We investigate pitfalls in this approach, discovering the issue of attribute amplification, where unrelated attributes are spuriously affected during interventions, leading to biases across protected characteristics and disease status. We show that attribute amplification is caused by …

abstract arxiv causal classifiers counterfactual cs.ai cs.cv generative image image generation images imaging issue look medical medical imaging modelling queries type work