2018 FALL LC MAGAZINE

Machine learning has come a long way since that early checkers game. Today, nearly 60 years after Samuel named it, machine learning allows self-driving cars to take passengers wherever they like. It’s what tells your smartphone to notify you about traffic conditions at the time you usually leave work — without your asking for it. It’s changing much about the way we live. In the laboratory of Christopher Kulp, Ph.D., professor of physics, students delve into the realm of machine learning to gain valuable research experience on the cutting edge of science. For example, Andrew Chase ’18, a physics major and astronomy minor, joined the lab to learn more about machine learning. Although the work was at first exploratory, he soon discovered that he may be able to determine a better way to distinguish exoplanets from false positives in observations made by the Kepler Space Telescope. Chase has been working with data plots that chart the various properties of exoplanets, including density, size, and orbital radius. Machine learning algorithms are then used to cluster the exoplanet data in ways that help him determine whether the observation was in fact a planet, or a false positive. “The planets I observed were found by looking at how a star’s brightness changes as a planet passes in front of the star. I also looked at how other astronomical events can mimic a planet,” said Chase. Through his work, the algorithm is now able to recognize whether an observation is a confirmed or an unconfirmed planet with 80 percent accuracy — a capability that could help astronomers with their observations. “This work is really relevant to graduate-level studies and will prepare me well to pursue a doctorate, or a career in physics,” he said. “I personally benefitted greatly from research as an undergraduate, and I place great importance on it for my students now,” said Kulp. “I work with students in a lab setting to give them a glimpse of how physics works at a professional level.” From Exoplanets to Fuel Efficiency oined by Arthur Samuel in 1959, the term “machine learning” is a form of artificial intelligence where computers progressively improve performance on a specific task, such as identifying patterns in data, without being explicitly programmed to do so. A pioneer in video gaming, Samuel’s checkers-playing program was one of the first self-learning programs ever created. C FROM Chase 14 LYCOMING COLLEGE 2018 FALL MAGAZINE