DARPA seeks ReMath submissions
On August 20, the Defense Advanced Research Projects Agency (DARPA) issued a call for submissions for the Recovery of Symbolic Mathematics from Code (ReMath) program. Submissions are due by 12:00 p.m. Eastern on September 18.
DARPA is issuing an Artificial Intelligence Exploration (AIE) Opportunity inviting submissions of innovative basic or applied research concepts in the technical domain of recovering symbolic mathematical expressions instantiated in software’s code. This AIE Opportunity is issued under the Program Announcement for AIE, DARPA-PA-20-02-01. All awards will be made in the form of an Other Transaction (OT) for prototype project. The total award value for the combined Phase 1 base and Phase 2 option is limited to $1,000,000. This total award value includes Government funding and performer cost share, as required or if proposed.
The ReMath AI (Artificial Intelligence) Exploration program will discover whether a combination of recent advances in AI techniques, such as neural machine translation, sequence-to-sequence encoders, etc., can effectively recover mathematical structures implemented in software in natural mathematical forms of expression. These techniques will improve the understanding of complex software and may enable future methods for analysis and testing of cyber physical systems.
ReMath proposals will identify and address challenges to application of AI methods, including (but not limited to) overcoming any potential scarcity of training data and developing novel effective representations that capture how mathematical symbolic units such as mathematical functions, equations, and formulas used by subject matter experts (SMEs) to describe and communicate about the domain concepts (“domain communication units”) are instantiated in source and binary code. In particular, proposers should discuss effective ways of automatically augmenting existing training sources or corpora, or automatically generating new training corpora, and should discuss problem representations that take advantage of expert analyses’ outcomes or captured domain SME work flows when understanding a system (if available).
Full information is available here.