A team of U of T computer scientists led by PhD student Sejal Bhalla has designed software that uses deep learning algorithms to decipher changes in vocal characteristics that indicate the lung condition of patients with chronic obstructive pulmonary disease.