“What we do is perform virtual treatment on virtual patients. The goal is to personalize medicine.”

That’s how B.J. Fregly described the impact of data science on his research as a mechanical engineer, one of many disciplines represented at the second-annual Rice Data Science Conference. The event was hosted Oct. 8-9 by the Ken Kennedy Institute for Information Technology, and drew some 400 leaders from industry and academia to the Bioscience Research Collaborative at Rice University.

“By definition, we are crossing boundaries, bringing together clinicians and engineers,” said Fregly, professor of mechanical engineering (MECH) and bioengineering at Rice, and CPRIT Scholar in Cancer Research, while taking part in a panel discussion titled “Opportunities and Challenges with Applying Data Science to Human Performance, Injury Prevention and Rehabilitation.”

Jan Odegard, executive director of the Ken Kennedy Institute, said the unofficial theme of the conference was “How Does Rice Plug into Houston?” The lineup of speakers included numerous Rice faculty and alumni, and representatives of Houston institutions.

In his introductory remarks, Reginald DesRoches, the William and Stephanie Sick Dean of Engineering at Rice, said of data science, “I can’t imagine a more timely and important topic. Processing, visualizing, and transforming data is fundamentally shifting strategic decision-making and core business operations in nearly every business environment.”

DesRoches, a professor of civil and environmental engineering and of MECH, related data science to his own work in resilient infrastructure: “We are now realizing that we can learn much more about disaster resilience by gathering data from the structure and systems minutes and days after an earthquake, than we can from testing.”

In his lab, Fregly has been working to develop personalized computer models of individual patients, to simulate various treatment options and identify new ones. His goal is to design clinical interventions to optimize post-treatment function for individual patients with such conditions as Parkinson’s disease, stroke, spinal-cord injuries, osteoporosis and cancer.

“I rely on data, lots of it: muscle strength, muscle electrical activity, joint contact force data, dynamic and static imaging, motion data. I need enormous quantities of information to build accurate models of individual human patients,” Fregly said.

Joining Fregly on the health and medicine panel was Marcia O’Malley, the Stanley C. Moore Professor of MECH at Rice and director of the Mechatronics and Haptic Interfaces Lab. “We leverage data to enhance human and robotic performance,” she said. “For instance, we need enormous quantities of motion data in order to develop human-machine interfaces for human-assisted movement and the modeling of human-robot interactions.”

Also on the panel was a frequent collaborator with O’Malley and Fregly, Dr. Gerard E. Francisco, the Chief Medical Officer for the TIRR Memorial Hermann/Memorial Hermann Rehabilitation Network.

“As physicians, we have not been trained to look at big numbers. It has been a process of education. Now we have learned to ask, what can Big Data do for our patients? We can tell practitioners they shouldn’t be intimidated by big numbers,” Francisco said.

Another panel member was John De Witt, the senior biomechanist at NASA’s Johnson Space Center and former performance coach with the Houston Dynamo Academy and other sports teams. His job focuses on preparing astronauts for planned missions to the Moon and Mars. “We are concerned with such questions as what is the optimal amount of exercise for an astronaut? We have the sensors to monitor so many functions,” he said. “How can we deal with all of this information?”

The Human Performance panel was moderated by Chris Culbert, the chief technologist at the Johnson Space Center, and also included Michael Watts, the performance coach and sport scientist for Under Armour.

Natalie Berestosvky, a data scientist with Anadarko Petroleum’s Advanced Analytics and Emerging Technology team, earned her M.S. and Ph.D. in computer science from Rice, where she worked in the research team headed by Luay Nakhleh. Her talk was titled “When Data Science and Geoscience Come Together.”

“When we deal with a previously developed drilling area, the well logs are not always optimal. The accuracy of log interpretation depends on the time spent on data quality control and on conditioning. The point is to identify bad data and ignore it,” she said.

Xaq Pitkow, assistant professor of electrical and computer engineering at Rice, and of neuroscience at Baylor College of Medicine, spoke on “Data Science for Studying the Data Scientist in Your Head.”

“We all go around and experience the world. The brain is matched to the world,” said Pitkow, who cautioned against “model bias” in the human brain and in algorithms with limited flexibility to discern the true signal from a dataset.

Akane Sano, assistant professor of electrical and computer engineering at Rice, discussed her research, which focuses on human sensing, data analysis and application development for health and wellbeing. She works on measuring, predicting and helping reduce stress, and improving general mental health, sleep and performance.

“Can we predict mood and stress for tomorrow using today’s data? Our goal is to design an embodied intelligent assistant to enhance cognitive performance. We have already looked at this for students and shift workers,” Sano said.

The conference featured two pre-conference workshops, seven plenary presentations, four parallel sessions and the presentation of 40 student posters. In addition to Odegard and Keith Cooper, the L. John and Ann H. Doerr Professor in Computational Engineering at Rice, who served as co-chairs, the workshop conference program committee included Natalie Berestovsky, Anadarko Petroleum; Alena Crivello, Chevron; Scott Ferguson, HEDS Group; Roy Keyes, Houston Data Science Group; Scott Morton, Rice; Craig Rusin, Baylor College of Medicine; Jim Ward, Two Sigma; and Yan Xu, Houston Machine Learning Meetup.

Rice Engineering: https://engineering.rice.edu/data_science_conference