Professor of Bioinformatics at Indiana University South Bend whose research spans computational biology, genomics, machine learning, and artificial intelligence. He has worked at the intersection of AI and biology since the early 1990s, applying artificial neural networks to the prediction and interpretation of biological sequence signals.
A distinctive aspect of his early work was the development of the Calliper randomization method for interpreting neural networks. The method systematically perturbs input features by introducing noise and measures the resulting change in model performance, thereby revealing which features are important to the network's learned representation. This approach conceptually anticipates the perturbation and ablation methods now widely used in explainable AI and in interpreting large language models, where components or information are selectively altered or removed to determine their contribution to model behavior.
His broader research encompasses sequence-dependent DNA structure, transcriptional regulation, alternative splicing, gene-expression analysis, cancer transcriptomics, and higher-order genome organization. His current research extends this long-standing interest in interpretable machine learning by integrating artificial intelligence with genomic, transcriptomic, and genome-structure data to uncover molecular mechanisms of gene regulation and disease. He is also committed to training students at the interdisciplinary interface of biology, bioinformatics, artificial intelligence, mathematics, and computer science.
Ali Movasati