IBM’s Watson schools Buckeyes on cognitive computing

Posted: 

img_3859_web.jpg
Students in the Knowledge-Based Systems course work in teams to develop software apps with guidance from Prof. Eric Fosler-Lussier (green sweater) and IBM experts like Stephen Boxwell (left).
Students in Eric Fosler-Lussier’s revamped capstone experience course are learning about knowledge-based systems in a completely new way. For the first time, the course challenges students to build software applications fueled by one of IBM’s most prized innovations: Watson. 

That’s right, the cognitive computer technology that bested two Jeopardy! champions in 2011 is now helping school the next generation of cognitive computer scientists.

The Ohio State University is one of just ten top technology schools chosen by IBM to launch the cognitive computing courses. Co-designed by the Watson Group and leading academic experts, the courses aim to empower students with the technical skills and hands-on learning required to develop new cognitive computing applications.

Open to both undergraduate and graduate students, the course challenges students to choose an industry and develop apps that help people make better decisions about a particular topic. After choosing to focus on various aspects of campus life at Ohio State, from student life to health services, the class split up into teams. Although Watson served as the backend technology for their apps, each team must build an interface that makes it easy for end users to understand what the system can do.

Computer Science and Engineering Associate Professor Eric Fosler-Lussier said the revised Knowledge-Base Systems course complements the College of Engineering’s extensive offerings in natural language processing, data mining and software engineering. 

“It’s an exciting opportunity for students across many disciplines, including majors in our new data analytics program,” he said. “With Watson as the ultimate virtual capstone team member, students are empowered to explore cognitive computing in depth, while promoting team design and lifelong learning principles.”

“The process of envisioning an idea, gathering requirements from people who would like to see it go live, and building upon it is very exciting,” said Chaitanya Shivade, a computer science and engineering graduate student who is enrolled in the course. “Being a PhD student, I am familiar with the research side of R&D. This course is teaching me the development side of it.” 

Unlike the version of Watson that appeared on Jeopardy!, what these Buckeye engineers are working with doesn’t come with a pre-compiled knowledge base. The system knows previously learned language relationships, but doesn’t actually contain any knowledge.

After working to enter relevant data into Watson, the teams must train it to make better associations.

“Part of what I’m hoping students get out of this class is understanding how machine learning algorithms can incrementally learn better and better answers to questions through continual training,” said Fosler-Lussier.

The addition of Watson also adds a natural language processing component to the course. Teams must design applications that ask a natural language query to the backend, while remaining as simple for the end user as possible. Middleware code must be written to process the answer and do the application logic in order for that to happen, Fosler-Lussier explained. 

Using Watson also requires students to work with significant text collections, which pose a unique challenge. 

“When we have gobs and gobs of text, how do we actually make sense of it and lead people to the text that helps them make decisions better?” Fosler-Lussier said. “This is really a big issue that a lot of computer scientists—especially in areas like computational humanities and biomedical informatics—are going to come up against.”

In order to boost students’ entrepreneurial skills, each team must create a business plan for their Watson apps. The goal is to enable students to launch their own companies in the future.

img_3846_web.jpg
Undergraduate students Ajda Gokcen (left) and Sarah Lee work to create a software app powered by IBM
“Project-based, competitive, language-heavy, and a chance to see and be seen by those in the industry—this class is a great opportunity all around,” said undergraduate student Ajda Gokcen. “I've learned so much about the practical aspects of making a real-world, industry-level application since starting this project.”

The students also work on technology teams to explore additional tools that might help them complete their projects. They report back to the class via oral presentations and online documentation. Researching topics from web APIs to natural language processing allows students to learn broadly about multiple technologies while gaining a deeper understanding in one area, Fosler-Lussier explained.

As a capstone experience course, the class also promotes teamwork, communications skills, planning and documentation to give students a realistic design-build project experience. 

“We don’t tell them how to do this project; they have to figure out how to get from A to B,” Fosler-Lussier said. “As the class facilitator, my job is to help them along the way.”

The IBM-Watson partnership is part of an ongoing effort to expand and strengthen student skills and understanding of big data and analytics in order to meet the growing demand for highly skilled analytics workers. The explosion of data-driven content has sparked a new wave of career opportunities for today’s college students, from business analytics professionals to chief data officers. Research firm Gartner, Inc., predicts that 4.4 million IT jobs will be created to support Big Data by 2015.

The collaboration is one of the latest developments in the continued partnership between IBM and The Ohio State University. Other joint efforts include establishing the IBM Client Center for Advanced Analytics in 2012, a first-of-its-kind partnership bringing together business, industry, and academic experts to create a hub for advanced analytics solutions. IBM is also a partner in the development of new business and technology curricula to help students gain the latest skills in analytics and prepare for high value careers. 

Written by Candi Clevenger, College of Engineering Communications, clevenger.87@osu.edu 

Category: Students