SONIC will be guided by the following mission: To enable equivalent scaling in beyond-CMOS nanoscale fabrics by embracing their statistical attributes within statistical-inference-based applications, architectures, and circuits, to achieve unprecedented levels of robustness and energy efficiency.
SONIC will achieve that mission by structuring its research agenda to employ three top-down systems-driven themes (Themes 1, 2, and 3) complemented by an experimental bottom-up hybrid devices and circuits theme (Theme 4).
Theme 1 (Stochastic Information Processing Systems) employs a unique, communications-inspired approach to nanoscale information processing and brings together the three seminal inventions of the information age (computing, communications, and semiconductors) to effect a fundamental transformation of the discipline of computing, taking it from a von Neumann manipulation of bits to an integrated processing of information.
Theme 2 (Information-based Mixed-signal Interfaces) emphasizes information as the metric of interest, rather than waveform preservation, in order to discover a process-scalable foundation for analog mixed-signal circuits and systems, and place analog circuits on par with their digital counterparts in terms of scaling.
Theme 3 (Cognitive Information Processing Systems) takes a unique approach of engaging neuroscientists, system architects, and circuit/device researchers, and has the potential to overcome the limitations of existing neuromorphic techniques to approach the brain’s energy efficiency and robustness by employing the power of higher-level learning, cognition, and statistical inference.
Theme 4 (Nanofunctions and Nanoprimitives) focuses on design of nanofunctions and nanoprimitives (rather than pure device research) and will accelerate the deployment of beyond-CMOS nanoscale fabrics.
Last Year2 Research Publications5 Patent Applications
Since Inception4 Research Themes11 Universities196 Students33 Faculty Researchers43 Liaison Personnel1,074 Research Publications5 Patent Applications4 Patents Granted