Amoeba-inspired computing architecture implemented using charge dynamics in parallel capacitance network

Kasai, Seiya; Aono, Masashi; Naruse, Makoto

doi:10.1063/1.4826143

Cited by 18 publications

(16 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A promising candidate of such a computing platform is an analogue electronic circuit. In fact, some authors have explored the use of the amoeba-inspired electronic circuits for tackling the constraint satisfaction problem [33], satisfiability problem [34], analogue-to-digital conversion [35] and finding walking manoeuvre of a multi-legged robot [36]. Such an approach to exploit physical parallelism may develop a novel analogue computing paradigm, providing powerful approximation methods for solving complex optimization problems appearing in a wide spectrum of real-world applications.…”

Section: Discussionmentioning

confidence: 99%

Remarkable problem-solving ability of unicellular amoeboid organism and its mechanism

et al. 2018

Self Cite

View full text Add to dashboard Cite

Choosing a better move correctly and quickly is a fundamental skill of living organisms that corresponds to solving a computationally demanding problem. A unicellular plasmodium of Physarum polycephalum searches for a solution to the travelling salesman problem (TSP) by changing its shape to minimize the risk of being exposed to aversive light stimuli. In our previous studies, we reported the results on the eight-city TSP solution. In this study, we show that the time taken by plasmodium to find a reasonably high-quality TSP solution grows linearly as the problem size increases from four to eight. Interestingly, the quality of the solution does not degrade despite the explosive expansion of the search space. Formulating a computational model, we show that the linear-time solution can be achieved if the intrinsic dynamics could allocate intracellular resources to grow the plasmodium terminals with a constant rate, even while responding to the stimuli. These results may lead to the development of novel analogue computers enabling approximate solutions of complex optimization problems in linear time.

show abstract

Section: Discussionmentioning

confidence: 99%

Remarkable problem-solving ability of unicellular amoeboid organism and its mechanism

et al. 2018

Self Cite

View full text Add to dashboard Cite

show abstract

“…We design SAPS using AQFP logic that can solve a small-scale NOR problem (N = 4). Figure 3(a) shows a schematic of SAPS, which is based on the amoeba-inspired semiconductor circuit demonstrated by Kasai et al [33] The AQFP logic gates in Fig. 3(a) are based on the buffer shown in Fig.…”

Section: Superconductor Amoeba-inspired Problem Solver (Saps)mentioning

confidence: 99%

“…3(d), which illustrates SAPS for N = 8. While the previously reported amoeba-inspired circuit[33] used circuit parameter variations to avoid deadlocked states, SAPS utilizes thermal fluctuation to avoid deadlocked states, which enables multiple solutions independent of initial states to be found, as will be shown later. SAPS includes 12 Josephson junctions per variable, excluding the peripheral circuits for readout.…”

mentioning

confidence: 99%

Superconductor Amoeba-Inspired Problem Solvers for Combinatorial Optimization

2019

Self Cite

View full text Add to dashboard Cite

Adiabatic quantum-flux-parametron (AQFP) logic is an energy-efficient superconductor logic family; the energy dissipation of an AQFP gate can be arbitrarily reduced through adiabatic switching. In addition to high energy efficiency, AQFP logic has the advantage that it can easily introduce stochastic processes by exploiting naturally occurring thermal fluctuations. In this paper, we propose using AQFP logic to implement an amoeba-inspired problem solver (APS), which is a stochastic local search method to explore solutions to combinatorial optimization problems such as the Boolean satisfiability problem (SAT). We designed a superconductor amoeba-inspired problem solver (SAPS) using AQFP logic, which finds solutions to a simple logical constraint satisfaction problem in the manner of APS, and fabricated it using a Nb integrated circuit fabrication process. Experimental results showed that the probability distribution of the stochastic processes in AQFP logic can be

show abstract

“…The amoeba-based solution searching described above utilizes spatiotemporal and probabilistic dynamics, indicating the possibility of realizing equivalent dynamics in solid-state materials [9,18]. In this study, we deal with optical energy transfer among multiple quantum dots as a concrete, quantitatively describable platform [9,10,16,[23][24][25].…”

Section: Photoexcitation Transfer Modelingmentioning

confidence: 99%

“…Naruse et al also demonstrated solving MAB by using single photons emitted from a nanodiamond [17]. Meanwhile, Kasai et al succeeded experimentally in solving CSP based on solid-state electrical circuits on the basis of spatiotemporal dynamics involving noise sources [18]. Recently, Aono et al examined nanoarchitecture for an SAT solver based on Gallium arsenide (GaAs) nanowires [19].…”

Section: Introductionmentioning

confidence: 99%

Category Theory Approach to Solution Searching Based on Photoexcitation Transfer Dynamics

et al. 2017

Self Cite

View full text Add to dashboard Cite

Solution searching that accompanies combinatorial explosion is one of the most important issues in the age of artificial intelligence. Natural intelligence, which exploits natural processes for intelligent functions, is expected to help resolve or alleviate the difficulties of conventional computing paradigms and technologies. In fact, we have shown that a single-celled organism such as an amoeba can solve constraint satisfaction problems and related optimization problems as well as demonstrate experimental systems based on non-organic systems such as optical energy transfer involving near-field interactions. However, the fundamental mechanisms and limitations behind solution searching based on natural processes have not yet been understood. Herein, we present a theoretical background of solution searching based on optical excitation transfer from a category-theoretic standpoint. One important indication inspired by the category theory is that the satisfaction of short exact sequences is critical for an adequate computational operation that determines the flow of time for the system and is termed as "short-exact-sequence-based time." In addition, the octahedral and braid structures known in triangulated categories provide a clear understanding of the underlying mechanisms, including a quantitative indication of the difficulties of obtaining solutions based on homology dimension. This study contributes to providing a fundamental background of natural intelligence.

show abstract

Amoeba-inspired computing architecture implemented using charge dynamics in parallel capacitance network

Cited by 18 publications

References 20 publications

Remarkable problem-solving ability of unicellular amoeboid organism and its mechanism

Remarkable problem-solving ability of unicellular amoeboid organism and its mechanism

Superconductor Amoeba-Inspired Problem Solvers for Combinatorial Optimization

Category Theory Approach to Solution Searching Based on Photoexcitation Transfer Dynamics

Contact Info

Product

Resources

About