At the Pacific Northwest National Laboratory, Tino Wells is working on machine learning, a form of artificial intelligence, or AI, that involves the development of computer programs that can access data and use it to learn for themselves. The know-how on researching and creating those computer programs is the skill set the national lab values in Wells.
In his senior year at the University of Hawai‘i at Hilo, astronomy and physics student Tino Wells had his sights on graduate school. He stayed busy finishing up his studies and research projects, and graduated in the spring of 2019 with a bachelor of science in astronomy and bachelor of arts in physics, cum laude on both counts. And then a national research institution of the United States Department of Energy offered him a long-term job.
Wells is now working as a research associate in deep learning, machine learning, and cheminformatics at the Biological Sciences Division of the Pacific Northwest National Laboratory (PNNL). The research institution, with its main campus located in Richland, Wa., is managed by the Department of Energy’s Office of Science.
As an undergraduate at UH Hilo, Wells was already an impressive scholar and researcher. When a senior, he conducted research exploring galaxies and rare stars. He was awarded two research opportunities to conduct inquiries that took him from UH Hilo to Notre Dame, building his skills and his résumé along the way. He also spent a year working with Kathy Cooksey, an associate professor of physics and astronomy at UH Hilo, on a project titled, “Non-parametric Clustering Analysis: Classifying Large-Scale Gaseous Structures in Circumgalactic Media to Establish Connections to Parent Galaxies,” studying the chemical signature in the gaseous regions around galaxies.
- Learn more about Wells’s research while at UH Hilo: UH Hilo astronomy & physics undergraduate Tino Wells researches galaxies and rare stars (UH Hilo Stories, Nov. 27, 2018).
It was Cooksey who introduced Wells to machine learning, a form of artificial intelligence, or AI, that involves the development of computer programs that can access data and use it to learn for themselves—the PNNL website describes their use of machine learning as artificial intelligence applied to scientific problems. And the know-how on researching and creating those computer programs is the skill set the national lab values in Wells.
“I am using a lot of the skills in my current role [that] I acquired studying astronomy/physics,” Wells explains in an email. “Mainly coding, and some quantum and atomic physics from formal courses. In addition, and debatably just as important, I couldn’t be in this position without the undergraduate research I did with Dr. Cooksey in machine learning. This started my journey in machine learning, and as this is the title of [my current] position, I’m sure you can infer its importance.”
At the Pacific Northwest National Laboratory, Wells along with several others are working under Ryan Renslow, a chemical engineer. According to Renslow’s bio on the lab’s website, his research focuses on the relationship between microbial community structure and function, the emergence of higher order properties in multispecies communities, and the identification of novel metabolites in these communities. To achieve this, he uses computational and mathematical tools, as well as advanced imaging techniques. This type of research has direct applications in a wide array of fields, including human health, energy production, ecology, national defense, machine vision, and biotechnology.
Wells is in charge of creating and developing an automated pipeline converting density-functional theory—for example, quantum chemical—data into an array of two-dimensional nuclear magnetic resonance (NMR) spectra via spin-dynamics simulations. “These spectra will be used for a multi-channel convolutional neural network and will eventually be integrated into our pre-existing model,” Wells explains. “This includes scripting in a couple of languages and frameworks. My project will be integrated into a powerful generative learning model called DarkChem: a variational autoencoder, developed by PNNL and those in our team, specifically focused on numerical and latent representations of molecular structures.”
Wells says the most surprising thing he’s discovered on the job so far is the extent to which machine learning is already integrated into chemistry. “Knowing machine learning can be a very powerful tool to many disciplines,” he says. “I find it truly fascinating to be directly involved with [the] integrative process.”
Wells credits his studies and research at UH Hilo with laying the foundation for his work at the national laboratory.
“I really could not be in the position I’m in today without the courses, faculty, and additional opportunities UH Hilo has provided,” he says. “The courses were essential, and working one-on-one with faculty was invaluable. The sheer amount of resources and opportunities for students at UH Hilo allows students to find exactly what fits for them; mine was machine learning.”
Story by Susan Enright, a public information specialist for the Office of the Chancellor and editor of UH Hilo Stories. She received her bachelor of arts in English and certificate in women’s studies from UH Hilo.