Deep-learning findings

Samuel Ehgartner ’24 and Ismail Jouny, professor and department head of electrical and computer engineering, published a scientific paper on deep-learning algorithms and delivered the findings during an engineering conference at Columbia University in October.

Ehgartner, who is studying electrical and computer engineering, started the project a few weeks after his spring 2023 semester. “We designed this new method of combining multiple algorithms together to make them more reliable and accurate,” Ehgartner says, explaining that they tested it with a mock security system.

Once the paper was complete, Ehgartner began looking for additional applications from this kind of deep learning. He found, for example, that when he took a dataset of mammography images, the algorithm allowed the computer to quickly compare images for benign and malignant tumors. Another application was showing the computer several satellite images of huge storm systems; with that sort of training data, the technology could test and possibly predict, say, the direction and severity of a storm in real time.

“I’m extremely grateful for the research opportunities that are at Lafayette and love the fact that the professors are very accessible,” says Ehgartner, who wants to ultimately pursue a job in the aerospace industry.