as Newsletter fall 2018

10 COLLEGE OF ARTS & SCIENCES NEWSLETTER We produce and encounter huge amounts of data every day. The photo of your lunch that you posted on Instagram? It’s now data. The credit card transaction that paid for your meal? That’s data as well. As you drove back to work from the restaurant, the traffic camera that filmed your car stored your movements as data, too. This data can prove useful. (Perhaps transportation engineers would like to know how to improve traffic at that intersection you drove through after lunch, for instance.) Analyzing and processing data to make predictions and gain insights requires computers with adequate processing power, but also a human being to tell it what to look for. “Data is unstructured and messy,” says K.C. Santosh. Ph.D., assistant professor of computer science. “It’s the computer scientist’s job to take that mess and turn it into something we can understand. We call that data mining and information retrieval.” Santosh works with colleagues throughout the world researching and developing ways to take data in all of its forms and turn it into useful information. Some areas that he has worked in extensively include computer vision, artificial intelligence, image analysis, pattern recognition and machine learning. Under the umbrella of artificial intelligence (AI)—Santosh prefers the term “close-to-human intelligence”—computer scientists help train computers to think in the same way as a human brain, only faster and error-free. “Humans can get tired and make mistakes,” he says. “AI tools don’t get tired.” Humans can also be in short supply. One of Santosh’s projects involves using a program to screen chest X-rays of HIV-positive patients in rural Kenya for evidence of tuberculosis and other pulmonary abnormalities. “We trained the machines using real doctors for several days with several different samples. It takes a lot of repetition,” he says. “And we took this machine to resource- constrained areas where they don’t have doctors or radiologists.” This type of intelligent data analysis is a tool in need of an application, Santosh adds. “We have many high-level applications in biology, biomedical engineering, physics, and chemistry, for example, but we need experts to tell us the semantics of the data before we start engineering,” he says. “What is the input? What should be processed? What do you want to have as an output?” Machine learning involves training computers to see patterns, detect deviations and make decisions based on the data they are receiving—learning and improving without being programmed to do a specific task. Working with colleagues in four different continents: Asia (Thailand, Japan, India) Africa (Kenya), Europe (France, Spain, Portugal, Sweden) and South America, Santosh’s projects in the realm of data mining and machine learning are numerous and varied. To counteract forgery, his team develops methods to quickly detect counterfeit logos on documents and forged signatures—even on documents with more than 12 different writing scripts commonly used in India. A project with the National Institutes of Health uses face recognition to help identify missing persons after a natural or man-made disaster. He is also developing a system that separates one voice from a recording with many voices and background sounds. Vehicular communication, fingerprint, palm-print and shoe-print verification, and online sketch and diagram analysis for architects are a few of his other projects. In addition, he is working on characterizing and classifying the Internet of Things (interconnected computing devices) in hospitals and universities through AI and machine learning tools. AI in robotics is another interest. Santosh, who was recently designated a senior member of the Institute of Electrical and Electronics Engineers (IEEE), writes and edits books and journals on the topic, organizes conferences and serves on boards and committees in his field. He also serves as an associate editor for the International Journal of Machine Learning and Cybernetics. He is an enthusiastic booster for his discipline and the benefits it can provide to all areas of the society. “So many projects can benefit from artificial intelligence and machine learning,” he says. “Data is everywhere.” Making Sense of Big Data K.C. Santosh, assistant professor of computer science, sees the potential of artificial intelligence in nearly all aspects of modern life, from healthcare to robotics. ‘Data is unstructured and messy. It’s the computer scientist’s job to take that and turn it into something we can understand.’ —K.C. Santosh