Science for Social Good - 2016 Projects
Open Discovery Platform for a Multiple Sclerosis Cure
Partner Organization: Accelerated Cure Project for Multiple Sclerosis
Motivation: Multiple sclerosis (MS) is a disabling disease of the central nervous system that disrupts the flow of information within the brain, and between the brain and body. The cause of MS is still unknown: scientists believe the disease is triggered by as-yet-unidentified environmental factor in a person who is genetically predisposed to respond. The progress, severity and specific symptoms of MS in any one person cannot yet be predicted. Analyzing heterogeneous data from clinical trials, clinical blood reports, gene-wide association studies and many other sources can lead to the discovery of factors associated with the disease. However, a single actor cannot hope to find such factors working alone; a platform for anyone in the world to collaboratively determine hypotheses to test and patterns to mine, with the assistance of cognitive technologies is a must.
Project Outcome: We developed a unique set of cognitive capabilities that understand the source code of a data analysis without any intervention by the user. This approach allows one to compare analyses based on an ontology, enabling a recommendation of similar and complementary analyses, visualization of the space of analyses, and so on, thereby accelerating discovery and knowledge sharing. Preliminary data analyses on factors that trigger multiple sclerosis, conducted using a large, heterogeneous MS database, have been analyzed through the cognitive capabilities.
Social Good Fellow: Evan Patterson, Department of Statistics, Stanford University
IBM Volunteers: Ioana Baldini Soares, Flavio Calmon, Huijing Jiang, Noi Sukaviriya, Justin Tyberg
Hunting Zika Virus with Machine Learning
Partner Organization: Cary Institute of Ecosystem Studies
Motivation: The flaviviruses are some of the most widespread viruses known. West Nile, Yellow Fever, and Dengue are a few of the best-documented examples in the Americas. The recent emergence of Zika virus outside of Africa has reiterated the need to discover what suites of correlated features of mosquitos and wild mammals combine to describe the most competent vectors and identify which wild primate species should be targeted for viral surveillance and management.
Project Outcome: We used physiological, behavioral, range, and social structure data of mammal species to develop a Bayesian predictive model of their status as reservoirs for different zoonotic diseases. The approach was trained on reservoir species from Africa to predict reservoir species in the Americas. The results will guide focal testing by disease ecologists in the field.
Social Good Fellow: Subhabrata Majumdar, School of Statistics, University of Minnesota
IBM Volunteers: Flavio Calmon, Raya Horesh, Abhishek Kumar, Elisa von Marschall, Adam Perer, Dennis Wei
What Works in Global Development?
Partner Organization: Clinton Global Initiative (CGI)
Motivation: CGI facilitates innovative solutions to the world’s most pressing problems through Commitments to Action that its members make and report measurable progress on. Best practices and lessons learned can be analyzed, captured, and disseminated from the corpus of 3,500 Commitments collected over 12 years. What best practices can be drawn from the Commitments that are relevant to others working in this space? How can work in this space be successfully scaled and replicated? What types of partnerships exist within these Commitments? What types of partnerships are most successful? What are key lessons learned as it relates to cross-sectoral partnerships? What common challenges exist across these Commitments?
Project Outcome: We developed a recommendation engine for Commitments by way of meta-analysis using natural language processing of Commitments and progress reports, along with network analysis of the members and partners involved. The recommendation engine can be used to guide new philanthropic investments and partnerships. We also developed a visualization that helps convey they advantages of bringing together different types of organizations to conduct social good projects.
Social Good Fellow: Hemank Lamba, School of Computer Science, Carnegie Mellon University
IBM Volunteers: Salman Baset, Anu Bhamidipaty, Mary Helander, Nizar Lethif, Joana Maria, Mauro Martino, Emily Ray, Moninder Singh, Daniel Weidele
Disseminate the Best Treatment for Diarrhea
Partner Organization: Clinton Health Access Initiative (CHAI), Inc.
Motivation: Diarrhea is the second leading killer of children under-5, responsible for more than 700,000 deaths globally each year. Zinc and oral rehydration salts (ORS) can prevent over 90 percent of diarrhea-related deaths and cost less than 50 cents per child; yet few children in need are receiving treatment. At the start of the program in 2011, an estimated 32 percent of children with diarrhea received ORS and less than one percent received the full recommended treatment. Instead, the majority of children would receive suboptimal products like antibiotics and antidiarrheals or nothing at all. In Nigeria, CHAI worked with the National Association of Proprietary Patent Medicine Dealers (NAPPMED), the trade union for over-the-counter medicine vendors, to train representatives to visit other medicine shops and teach their peers about ORS and zinc, with the goal of increasing their stocking of the treatments.
Project Outcome: We performed statistical analysis on CHAI’s program and found medicine shop owners who were visited by a representative were more likely to have correct knowledge of the best diarrhea treatments and were more likely to stock ORS and zinc. The results will help CHAI refine their programs around the world.
Social Good Fellow: Yumeng Tao, Department of Civil and Environmental Engineering, University of California, Irvine
IBM Volunteers: Debarun Bhattacharjya, Aliza Heching, Emily Ray, Moninder Singh, Aditya Vempaty
Real-Time Understanding of Humanitarian Crises
Partner Organization: ACAPS (Assessment Capacities Project)
Motivation: Over the last decade there has been notable improvement in the collection and dissemination of humanitarian information used to analyze the nature and magnitude for major crises around the world, and suggest better targeted response priorities. There has been healthy growth in the recognition of assessment and situation analysis as vital to humanitarian program design and monitoring. Through the last few years the systematic use of secondary data to inform and provide context for emergency programming has become a norm. At the same time, this has led to the emergence of analysis units and experts across thirty agencies and NGOs doing similar labor-intensive humanitarian assessment. These activities can be automated and improved using machine learning and information retrieval techniques.
Project Outcome: We developed an end-to-end system accessible through an API that performs focused web crawling to bring back articles only relevant to humanitarian crises, classifies them by type of disaster, and provides a faceted search interface to access the results using Watson technologies. The system improves the quality and depth of humanitarian analysis using data science techniques to take advantage of the increased volume and variety of available information.
Social Good Fellow: Kien Pham, Computer Science and Engineering Department, New York University
IBM Volunteers: Ioana Baldini Soares, Amit Dhurandhar, Arpith Jacob, Abhishek Kumar, Dmitry Malioutov, Prasanna Sattigeri, Maja Vukovic
How to Foster Innovation
Partner Organization: An International Non-Profit Foundation
Motivation: Innovation plays an essential role in the development of the modern global economy. It ranks among the most important of human activities, driving economic growth through the creation of job opportunities, new products and services, motivating cities, regions and countries to create environments that foster it to improve their competitiveness in local and global markets. Innovation is also a key component of sustainable development, and a means by which we uplift humanity. While innovation is easy to perceive, it is difficult to define, and consequently even more difficult to measure. In this project, we aim to do so and to furthermore understand the economic, sociological and anthropological drivers of innovation.
Project Outcome: Based on annual executive opinion surveys conducted by the partner organization, we constructed a time series of innovation scores per country and analyzed this data in conjunction with a large number of country-level metrics and indicator time series from the World Bank. The analysis yielded an open innovation index that infers the innovativeness of all countries of the world in a completely data-driven way and presents the factors associated with this inference. World leaders can use the results to understand what investments should be made to make their countries more innovative.
Social Good Fellow: Caitlin Kuhlman, Department of Computer Science, Worcester Polytechnic Institute
IBM Volunteers: Lei Cao, Aurelie Lozano, Karthikeyan Natesan Ramamurthy, Chandra Reddy, Prasanna Sattigeri