IBM Research: Data Science for a Better World

Saška Mojsilović led the creation of the Ebola Open Data Jam in 2014. A community effort which helped to identify, inventory and classify all open data sources related to the Ebola outbreak, it provided governments, aid agencies and researchers with free and open access to valuable open data related to the epidemic on the platform EbolaData.org.

IBM Research scientist Kush Varshney

Kush Varshney is a data ambassador for DataKind, an organization that brings together leading data scientists with high impact social organizations to better analyze, model and visualize data in the service of humanity. He has led projects with NGO GiveDirectly and Simpa Networks. GiveDirectly delivers cash directly to extremely poor villagers in Kenya and Uganda.

Using a combination of volunteers, satellite imagery, image processing and a machine learning-based algorithm, the team trained a system to identify the poorest villages based on the proportion of thatch to metal roofed homes, a simple yet effective proxy for poverty. A resulting paper on the research was awarded the best social good paper at the 2014 KDD Conference.

IBM Fellow Saška Mojsilović

For Saška, an IBM Fellow (the company’s highest technical rank), putting her skills as a data scientist to societal issues is a professional responsibility. “There are so many disparities in the world from justice, to poverty, to hunger. We need to try to do something. If everyone did the best they could do – the world might go in a different direction. We have a responsibility as people and scientists. And we have a responsibility to raise new generations of data scientists to think and act the same.”

For Kush, an IBM Research scientist and member of Saška’s team, it’s also about following family example. “My grandfather was dedicated to many social causes in India throughout his life. He was an applied mathematician like me, but in his day, there was no data being collected in the social sector, so he used other means to make a difference. Today, I can use my expertise to create a better world.”

Today, their sense of purpose has led to the founding of IBM’s Social Good Fellowship Program. This new initiative invites graduate and post-doctoral students to apply their computer science and statistical skills to projects that address important social issues, such as health care, equality, the environment, sustainability, disaster response and more.

As co-directors of the program, Saška and Kush bring a tremendous amount of knowledge, relationships and experience. They’re what you might call hands-on philanthropic scientists, who will pair students and scholars with IBM scientists and partner organizations to tackle some of the biggest challenges in our world.

Open Data at a Tipping Point

The social good movement has taken root with many a corporation, entrepreneur and big thinker, with the simple aim of using technology to help create a better world. Data science is one increasingly important way that social good can be made possible and an entire community has grown around it, fueled in large part by the fact we are no longer constrained by data.

IBM Research scientists Raya Horesh (standing, far left), Dennis Wei (standing, center), Kush Varshney (sitting, right), and Saška Mojsilović (far right) with data science community members at the second Ebola Open Data Jam in 2015.

“Everything from Internet activity, satellite imagery, social media, health records, news, scientific publications, economic data, weather data, and government records is all at our fingertips, giving us an opportunity, and a responsibility, to change the world for the better using data science,” said Saška.

Two types of fellowships will be offered through the program. Starting this summer, IBM Research will bring on graduate students for a 12-week fellowship, and later in the year a postdoctoral scholar will be hired to complete a one-year fellowship. Each student will be partnered up with one or more IBM Research mentors who will assist and guide the data scientists throughout their term. Over time, Kush and Saška hope to expand the number of hires, partners, mentors, IBM labs, and scientific disciplines that make up the program.

What’s notable about the IBM fellowship is that it’s the first corporate program aimed exclusively at data science for social good, focused on advanced students with skills in computer or data science and analytics. IBM is now joining the data for good movement, which has so far been largely shaped by startups, non-profits, hackathons/challenges and university-based programs. Friends who are active players in this community, such as the University of Chicago and DataKind, shared many lessons with Saška and Kush that were instrumental in helping shape the creation of the IBM effort.

Key to the success of the program is scoping. Well before any student starts, the IBM team will have worked with partners to secure and curate the relevant data, lay out the objectives and agree on outcomes.

For Kush and Saška, this program means more than giving the next generation of data scientists projects with which to hone their skills and produce publishable results. It’s also helping them make an impact on humanity by addressing issues important to the general population. And then there’s another type of giving with the involvement of IBM mentors. “I can’t even explain the amount of enthusiasm we have received for this program. Everyone wants to participate,” Saška says.