Louis Casteilla

This study uses ML to classify regenerating vs. scarring mouse tissue images, finding models overfit to individual mice rather than biological features, but successfully extract relevant biological information when tasks align with data capabilities.

The paper introduces a visualization method to analyze SHAP feature interactions in ML models for biomedical applications, revealing hidden patterns of explainability.

This paper introduces discernibility as a metric to evaluate the clarity of explanations in machine learning models for medicine, finding that the dcor metric best captures user-perceived relationships between data and explanations.

This study shows opioids inhibit regeneration by blocking granulocyte ROS production in vertebrates, and this effect can be reversed with an opioid antagonist.