Deep learning perturbation models can outperform baselines on calibrated metrics

Summary

Research article on deep learning perturbation models demonstrating superior performance on calibrated metrics. No specific company, drug, or clinical development information provided.