This page is optimized for a taller screen. Please rotate your device or increase the size of your browser window.

Abt Expands Artificial Intelligence (AI) Capabilities

Data Science Fellowship Program Aims To Train 100 Staff in AI-related Skills

Rockville, Md. - Abt Global has launched an initiative to increase staff capacity to use artificial intelligence (AI) to streamline and improve analysis of vast amounts of data. The Abt Data Science Fellowship program aims to train 100 staff members by April 1, 2021, to use machine learning, predictive analytics and natural language processing in their subject matter areas and specialties. The specialties include environment, health, education, work force, criminal justice and housing.

In a deliberate departure from the male-dominated data-science field, a majority of Abt’s Data Science Fellowships will be awarded to women, adding needed diversity to this field and reflecting Abt’s fulltime U.S. employees, 65 percent of whom are women. 

“Federal agencies such as the Centers for Medicare and Medicaid Services (CMS) increasingly are exploring the use of such analytical capabilities to solve problems in new ways that go beyond what is possible with other analytic techniques,” said Christopher Spera, Ph.D., Division Vice President, Health and Environment. “In Abt’s work in partnership with CMS, for example, we are exploring the use of machine learning to help select comparison groups that maximize precision and improve estimates in a quasi-experimental evaluation.”

CMS recently issued an AI health data challenge to accelerate innovative solutions to improve predictions of health outcomes and the quality of care for patients.  While the defense industry has used it for some time, AI is now moving into social, health and economic policy research and decision-making.

One Abt Data Science Fellow’s project for the U.S. Department of Education’s What Works Clearinghouse, for example, involved a broad landscape review of research on college interventions to help students, which yielded around 50,000 abstracts. Developing a natural language processing algorithm to narrow the focus took several days, but running it took just seconds. The result was a comprehensive map of the different topics in college education intervention that researchers had investigated.

Another Abt Data Science Fellow created a dataset consisting of more than 100 journal article abstracts from Google Scholar. The Fellow applied natural language processing techniques to identify the key words in the dataset and sort the key words into the algorithm’s predicted topics. The algorithm enabled the Fellow to identify which articles have those words and which groups of articles share the same topics. This approach could help analysts recommend relevant articles for literature review for a library database of thousands of journal articles. 

The Data Science Fellows use Python to develop the algorithms and Jupyter Notebooks as the user interface to run the programming language, create and display their data visualizations and report their findings. Both are free. Abt also uses and contributes to free algorithms in open source libraries.

Abt already is using AI and machine learning for clients. The U.S. Department of Labor awarded Abt a $3 million contract in part to use AI to improve understanding of previous research and current web-based information on career pathways. Using machine learning to analyze 500 papers or 50 websites on the topic will be far faster than a manual review.

Abt also has used machine learning and image recognition to count mosquito eggs from a smartphone picture. We have used it to dig through vast numbers of open-ended survey responses to determine the main themes and individuals’ overall sentiment. And Abt has analyzed spatial data to predict the onset of droughts.

“Abt is merging subject matter expertise with AI to make qualitative and quantitative improvements in our analyses,” said Spera. “And over time we will have more and more people who can incorporate the ever-changing best practices in AI to help improve the lives of the world’s most vulnerable people.”

About Abt Global
Abt Global is an engine for social impact, dedicated to moving people from vulnerability to security.  Harnessing the power of data and our experts’ insights, we provide research, consulting and technical services globally in the areas of health, environmental and social policy, technology and international development.

Mary Maguire


Work With Us
Ready to change people's lives? We want to hear from you.
We do more than solve the challenges our clients have today. We collaborate to solve the challenges of tomorrow.