A Georgetown University-based public policy think tank has found some differences and overlaps in the research agenda at six U.S. technology companies that pursue artificial intelligence and machine learning programs.
The Center for Security and Emerging Technology said it analyzed an aggregate dataset of 270,802 AI-related publications from IBM, Microsoft, Amazon, Google, Apple and Facebook in the past decade to gain insight into the primary research topic areas at corporate laboratories.
The top five topics listed in the report by publication count are robotics and grasping, modeling of 3D shapes, images, sparse matrices and representation, and fault/failure diagnosis.
CSET noted that the companies' differing areas of focus should be considered by policymakers when they craft an assessment framework for technological competitiveness and a government research investment strategy to align national policy with private-sector agendas around the emerging technology.
Authors of the study observed that the five technology firms do not prioritize work on federated learning and other machine learning areas that may help the U.S. address structural hurdles to AI adoption and advancement.
CSET data scientist Rebecca Gelles, research fellow Tim Hwang and research assistant Simon Rodriguez jointly authored the "Mapping Research Agendas in U.S. Corporate AI Laboratories" report.