Jane Edwards
The Defense Advanced Research Projects Agency has asked industry to provide information on mathematical frameworks and related methods in an effort to identify fundamental limitations in machine learning systems.
DARPA said Thursday the request for information for the Fundamental Limits of Learning program aims to find answers to several questions, including the efficiency of a learning algorithm to a specific problem and effects of noise and error on the training data.
“What’s lacking… is a fundamental theoretical framework for understanding the relationships among data, tasks, resources, and measures of performance — elements that would allow us to more efficiently teach tasks to machines and allow them to generalize their existing knowledge to new situations,” said Reza Ghanadan, program manager at DARPA.
“With Fun LoL we’re addressing how the quest for the ultimate learning machine can be measured and tracked in a systematic and principled way,” Ghanadan added.
DARPA said it expects to receive responses in various technical areas, including artificial intelligence, machine learning, cognitive science, information theory, statistics and computer science theory.
Responses to the RFI are due June 7, according to a notice posted on FedBizOpps.
