(1) PAPER TITLE
Universal Mechanical Polycomputation in Granular Matter
(2) AUTHORS
Atoosa Parsa
E434 Innovation Hall
University of Vermont
Burlington, VT 05405
802-557-7643
atoosa.parsa@uvm.edu
Sven Witthaus
The Faboratory at Yale University
9 Hillhouse Ave, ML 118
New Haven, CT 06511
203-432-5592
sven.witthaus@yale.edu
Nidhi Pashine
The Faboratory at Yale University
9 Hillhouse Ave, ML 118
New Haven, CT 06511
203-432-5592
nidhi.pashine@yale.edu
Corey S. O’Hern
The O’Hern Group at Yale University
9 Hillhouse Avenue
Mason Laboratory, Room 203
New Haven, CT 06511-6815
203-432-4258
corey.ohern@yale.edu
Rebecca Kramer-Bottiglio
The Faboratory at Yale University
9 Hillhouse Ave, ML 118
New Haven, CT 06511
203-432-5592
rebecca.kramer@yale.edu
Josh Bongard
E428 Innovation Hall
University of Vermont
Burlington, VT 05405
802-656-4665
josh.bongard@uvm.edu
(3) CORRESPONDING AUTHOR
Atoosa Parsa, atoosa.parsa@uvm.edu
(4) ABSTRACT
Unconventional computing devices are increasingly of interest as they can operate in environments hostile to silicon-based electronics, or compute in ways that traditional electronics cannot. Mechanical computers, wherein information processing is a material property emerging from the interaction of components with the environment, are one such class of devices. This information processing can be manifested in various physical substrates, one of which is granular matter. In a granular assembly, vibration can be treated as the information-bearing mode. This can be exploited to realize “polycomputing”: materials can be evolved such that a single grain within them can report the result of multiple logical operations simultaneously at different frequencies, without recourse to quantum effects. Here, we demonstrate the evolution of a material in which one grain acts simultaneously as two different NAND gates at two different frequencies. NAND gates are of interest as any logical operations can be built from them. Moreover, they are nonlinear thus demonstrating a step toward general-purpose, computationally dense mechanical computers. Polycomputation was found to be distributed across each evolved material, suggesting the material’s robustness. With recent advances in material sciences, hardware realization of these materials may eventually provide devices that challenge the computational density of traditional computers.
(5) LIST OF CLAIMED CRITERIA
A, D, G
(6) STATEMENTS
A: We believe the granular metamaterial designs that we have found would qualify as a patentable invention. There are examples of artificially designed materials that have been patented previously. The computational capabilities of our designs which enable superposition of multiple logical functions in one granular metamaterial, the high dimensional design space and the complexity of emerging patterns make the designs found by artificial evolution unique compared to previous work. Therefore, we are in the process of patenting our designs.
D: In recent years there has been growing research on embedding mechanical computation into the material. Despite recent advances, in none of those works is the computational unit automatically optimized to perform computation, let alone how best to densely pack computation in new ways into materials is explored. Therefore, we think that our result is valuable on its own as a new scientific discovery. Moreover, we are exploring the verification of our simulation results in physical hardware. It is possible that given the discrete nature of granular metamaterials compared to continuous media, crossing the reality gap may prove easier for former compared to the latter. Because different designs are currently just different placements of different types of particles on a predefined grid, we expect the fabrication process to be cheaper, faster and easier as well.
G: The non-intuitive nature of embedding computation into granular metamaterial is evidenced by the lack of obvious common patterns across the evolved materials that best embody the logic gates: each has unique ratios of soft/stiff particles and geometric patterns, and there is no obvious regularity or symmetry. This emphasizes the utility of automated design in this domain: designing a configuration of particles to behave as a logic gate is a rather difficult if not impossible task to accomplish without the aid of computer optimization.
(7) FULL CITATION OF THE PAPER
Parsa, A., Witthaus, S., Pashine, N., O'Hern, C. S., Kramer-Bottiglio, R., & Bongard, J. (2023). Universal Mechanical Polycomputation in Granular Matter. In Proceedings of the Genetic and Evolutionary Computation Conference, 2023, in press.
Parsa, A., Wang, D., O'Hern, C. S., Shattuck, M. D., Kramer-Bottiglio, R., & Bongard, J. (2022, July). Evolving programmable computational metamaterials. In Proceedings of the Genetic and Evolutionary Computation Conference (pp. 122-129).
(8) PRIZE BREAKDOWN STATEMENT
Any prize money, if any, is to be divided equally among the co-authors.
(9) REQUIRED STATEMENT INDICATING WHY THIS ENTRY COULD BE THE "BEST"
Metamaterials are an emerging class of engineered composite materials that exhibit properties different from their constituent materials and behave in ways not observed in nature. Most of the work in the field of metamaterial design have been essentially a human-driven process of trial and error to design materials with desired properties. Our work has two major aspects:
1. An evolutionary algorithm is used to solve the inverse design problem: granular metamaterials have been studied to a great extent in the material sciences literature. However, their high-dimensional design space and the unintuitive relationship between microstructure and desired macroscale behavior makes the inverse design problem formidable. In our research, for the first time, we showed the successful application of evolutionary algorithms for designing computational granular metamaterials.
2. Multiobjective optimization is able to evolve increasing amounts of computational potential into granular metamatetrials: to the best of our knowledge, our work is the first example of a single material design that can perform more than one logical function without any changes to the physical configuration after the fabrication, without recourse to quantum effects, and only by manipulation of the input frequencies. We showed how the input frequency is the key to regulate the functionality of the designed metamaterial and potentially increase its computational density. This opens up the door to numerous possibilities for packing more computational power into the material and possibly programming it to perform computation that it originally wasn’t designed to be able to perform.
(10) THE EVOLUTIONARY COMPUTATION TECHNIQUE USED
Genetic Algorithms (GA), Age-Fitness Pareto Optimization (AFPO)
(11) THE DATE OF PUBLICATION
Our recent paper has been accepted as a full paper at GECCO 2023. We have also published a previous related work at GECCO 2022.