Current rate of structured and unstructured data generation and the need for real-time data analytics requires radically different computational approaches that can operate in a massively parallel, scalable and energy efficient manner. In addition, certain classes of computational problems, i.e. combinatorial optimization problems, which have extensive applications in many real-world situations such as fault diagnosis, scheduling, resource allocation and even neural network training are fundamentally difficult to solve using the Boolean framework, the backbone of our current computational framework. This proposal aims to explore the potential of coupled oscillator networks made of living heart muscle cells, or bio-oscillators, as a collective computing fabric for solving computationally hard problems such as optimization, learning and inferences. New computing paradigms that can solve computationally hard problems efficiently using cell-based collective computing fabrics will transform how synthetic biocomputing circuits are built. In addition, the proposed research will lead to a better understanding of the electrical communication in muscle cell networks impacting potential future applications ranging from biorobotics to understanding and treating muscle disorders. The broader impacts will further be achieved through interdisciplinary student mentoring and education by the PIs with complementary backgrounds in Biology, Engineering and Computer Science, timely dissemination of key research outcomes via published papers and presentations, as well as proposed outreach activities including a workshop for middle school girls providing hands-on activities in STEM, mentoring of high school students, and a weekend long workshop each year for undergraduate and graduate students, called 'Biology Inspired Computing'. These activities will contribute towards educating a new cadre of students that will meet the future need of this emerging field. Attention will be paid to student recruitment from under-represented groups by participating in such recruitment programs at the PIs' institution.
Current designs that explore biological components for biocomputing leverages the information processing units of the cells, such as DNA, gene or protein circuitries, which are inherently slow (hours to days speed). Using electrically active cells that could individually operate in the hundred Hertz regime, and can be connected as networks to perform massively parallel tasks, can transform biocomputing and lead to novel ways of energy efficient information processing. The goal of this project is to explore the potential of electrically coupled oscillator networks made of living heart muscle cells to form a collective computing fabric for solving computationally hard problems such as optimization, learning and inference tasks. Heart muscle cells are electrically active components that can initiate and relay electrical signals without loss. More interestingly, they spontaneously beat (i.e. oscillate) at a stable pace, and when coupled with each other through ion fluxes, they synchronize to a locked, steady frequency. In this study, it is hypothesized that reconfigurable circuits fabricated using coupled heart muscle cell oscillators, or bio-oscillators, can be configured into functional continuous-time dynamical systems to solve computationally hard problems. Towards this end, state-of-art nanobiofabrication methods will be used to create bio-oscillators and to bi-directionally couple them through nanoporous ionic membranes for programmable computing. The design space will be explored for the geometry and size of the individual bio-oscillators, nanoporous membrane design for bio-oscillator connectivity through ion fluxes, as well as the network topology and fabrication feasibility of large bio-oscillator arrays. First, a pair of coupled bio-oscillators will be studied and computationally efficient compact models will be developed that accurately reflect the continuous time dynamics of the coupled bio-oscillators. Then, the design will be scaled up to larger network of coupled bio-oscillators and a feasibility study would be conducted for solving a prototypical hard optimization task such as vertex coloring of graph and graph partitioning. In addition to their inherent connectivity, scalability and parallel processing ability, the muscle cell-based approach proposed in this study requires minimal energy mostly in the form of sugar, and as such will be a low-energy alternative to current energy demanding traditional computing approaches for solving hard optimization problems using heuristics based digital computing techniques.
This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.
