DARPA is seeking innovative proposals for R&D of scalable, robust and power-efficient analog neural network architectures and circuits that could directly interface with the analog outputs of conventional sensors. ScAN systems will demonstrate inferencing capabilities of analog neural networks while eliminating the need for analog-to-digital converters at the raw sensor level. ScAN systems are also expected to demonstrate a three orders of magnitude power reduction over existing solutions.
ScAN will be a 54 month program:
Phase 1a (base) – 15 months – develop techniques to address TC1 and TC2 for realization of robust, accurate, and power efficient analog NN circuits at intermediate scales
Phase 1b (option) – 12 months – validation of techniques developed in Phase 1a
Phase 2 (option) – 27 months – develop and demonstrate techniques to scale analog NNs to large systems.
Abstracts are strongly encouraged but not mandatory and are due 8 July 2024. Full proposals are due 22 August 2024.
Document
Broad Agency Announcement – ScAN: HR001124S0022
