Nanofluidic logic with mechano-ionic memristive switches.

Neuromorphic systems are typically based on nanoscale electronic devices, but nature relies on ions for energy-efficient information processing. Nanofluidic memristive devices could thus potentially be used to construct electrolytic computers that mimic the brain down to its basic principles of operation. Here we report a nanofluidic device that is designed for circuit-scale in-memory processing. The device, which is fabricated using a scalable process, combines single-digit nanometric confinement and large entrance asymmetry and operates on the second timescale with a conductance ratio in the range of 9 to 60. In operando optical microscopy shows that the memory capabilities are due to the reversible formation of liquid blisters that modulate the conductance of the device. We use these mechano-ionic memristive switches to assemble logic circuits composed of two interactive devices and an ohmic resistor.

Self-averaging of digital memcomputing machines.

Digital memcomputing machines (DMMs) are a new class of computing machines that employ nonquantum dynamical systems with memory to solve combinatorial optimization problems. Here, we show that the time to solution (TTS) of DMMs follows an inverse Gaussian distribution, with the TTS self-averaging with increasing problem size, irrespective of the problem they solve. We provide both an analytical understanding of this phenomenon and numerical evidence by solving instances of the 3-SAT (satisfiability) problem. The self-averaging property of DMMs with problem size implies that they are increasingly insensitive to the detailed features of the instances they solve. This is in sharp contrast to traditional algorithms applied to the same problems, illustrating another advantage of this physics-based approach to computation.

Custodial Chiral Symmetry in a Su-Schrieffer-Heeger Electrical Circuit with Memory.

Custodial symmetries are common in the standard model of particle physics. They arise when quantum corrections to a parameter are proportional to the parameter itself. Here, we show that a custodial symmetry of the chiral type is also present in a classical Su-Schrieffer-Heeger (SSH) electrical circuit with memory. In the absence of memory, the SSH circuit supports a symmetry-protected topological edge state. Memory induces nonlinearities that break chiral symmetry explicitly and spread the state across the circuit. However, the resulting state is still protected against perturbations by the ensuing custodial chiral symmetry. These predictions can be verified experimentally and demonstrate the interplay between symmetry and memory.

Mode-assisted joint training of deep Boltzmann machines.

The deep extension of the restricted Boltzmann machine (RBM), known as the deep Boltzmann machine (DBM), is an expressive family of machine learning models which can serve as compact representations of complex probability distributions. However, jointly training DBMs in the unsupervised setting has proven to be a formidable task. A recent technique we have proposed, called mode-assisted training, has shown great success in improving the unsupervised training of RBMs. Here, we show that the performance gains of the mode-assisted training are even more dramatic for DBMs. In fact, DBMs jointly trained with the mode-assisted algorithm can represent the same data set with orders of magnitude lower number of total parameters compared to state-of-the-art training procedures and even with respect to RBMs, provided a fan-in network topology is also introduced. This substantial saving in number of parameters makes this training method very appealing also for hardware implementations.

Synaptic Plasticity in Memristive Artificial Synapses and Their Robustness Against Noisy Inputs.

Emerging brain-inspired neuromorphic computing paradigms require devices that can emulate the complete functionality of biological synapses upon different neuronal activities in order to process big data flows in an efficient and cognitive manner while being robust against any noisy input. The memristive device has been proposed as a promising candidate for emulating artificial synapses due to their complex multilevel and dynamical plastic behaviors. In this work, we exploit ultrastable analog BiFeO3 (BFO)-based memristive devices for experimentally demonstrating that BFO artificial synapses support various long-term plastic functions, i.e., spike timing-dependent plasticity (STDP), cycle number-dependent plasticity (CNDP), and spiking rate-dependent plasticity (SRDP). The study on the impact of electrical stimuli in terms of pulse width and amplitude on STDP behaviors shows that their learning windows possess a wide range of timescale configurability, which can be a function of applied waveform. Moreover, beyond SRDP, the systematical and comparative study on generalized frequency-dependent plasticity (FDP) is carried out, which reveals for the first time that the ratio modulation between pulse width and pulse interval time within one spike cycle can result in both synaptic potentiation and depression effect within the same firing frequency. The impact of intrinsic neuronal noise on the STDP function of a single BFO artificial synapse can be neglected because thermal noise is two orders of magnitude smaller than the writing voltage and because the cycle-to-cycle variation of the current-voltage characteristics of a single BFO artificial synapses is small. However, extrinsic voltage fluctuations, e.g., in neural networks, cause a noisy input into the artificial synapses of the neural network. Here, the impact of extrinsic neuronal noise on the STDP function of a single BFO artificial synapse is analyzed in order to understand the robustness of plastic behavior in memristive artificial synapses against extrinsic noisy input.

Directed percolation and numerical stability of simulations of digital memcomputing machines.

Digital memcomputing machines (DMMs) are a novel, non-Turing class of machines designed to solve combinatorial optimization problems. They can be physically realized with continuous-time, non-quantum dynamical systems with memory (time non-locality), whose ordinary differential equations (ODEs) can be numerically integrated on modern computers. Solutions of many hard problems have been reported by numerically integrating the ODEs of DMMs, showing substantial advantages over state-of-the-art solvers. To investigate the reasons behind the robustness and effectiveness of this method, we employ three explicit integration schemes (forward Euler, trapezoid, and Runge-Kutta fourth order) with a constant time step to solve 3-SAT instances with planted solutions. We show that (i) even if most of the trajectories in the phase space are destroyed by numerical noise, the solution can still be achieved; (ii) the forward Euler method, although having the largest numerical error, solves the instances in the least amount of function evaluations; and (iii) when increasing the integration time step, the system undergoes a "solvable-unsolvable transition" at a critical threshold, which needs to decay at most as a power law with the problem size, to control the numerical errors. To explain these results, we model the dynamical behavior of DMMs as directed percolation of the state trajectory in the phase space in the presence of noise. This viewpoint clarifies the reasons behind their numerical robustness and provides an analytical understanding of the solvable-unsolvable transition. These results land further support to the usefulness of DMMs in the solution of hard combinatorial optimization problems.

Thousand-fold Increase in Plasmonic Light Emission via Combined Electronic and Optical Excitations.

Surface plasmon enhanced processes and hot-carrier dynamics in plasmonic nanostructures are of great fundamental interest to reveal light-matter interactions at the nanoscale. Using plasmonic tunnel junctions as a platform supporting both electrically and optically excited localized surface plasmons, we report a much greater (over 1000× ) plasmonic light emission at upconverted photon energies under combined electro-optical excitation, compared with electrical or optical excitation separately. Two mechanisms compatible with the form of the observed spectra are interactions of plasmon-induced hot carriers and electronic anti-Stokes Raman scattering. Our measurement results are in excellent agreement with a theoretical model combining electro-optical generation of hot carriers through nonradiative plasmon excitation and hot-carrier relaxation. We also discuss the challenge of distinguishing relative contributions of hot carrier emission and the anti-Stokes electronic Raman process. This observed increase in above-threshold emission in plasmonic systems may open avenues in on-chip nanophotonic switching and hot-carrier photocatalysis.

Application of Floquet theory to dynamical systems with memory.

We extend the recently developed generalized Floquet theory [Phys. Rev. Lett. 110, 170602 (2013)] to systems with infinite memory, i.e., a dependence on the whole previous history. In particular, we show that a lower asymptotic bound exists for the Floquet exponents associated to such cases. As examples, we analyze the cases of an ideal 1D system, a Brownian particle, and a circuit resonator with an ideal transmission line. All these examples show the usefulness of this new approach to the study of dynamical systems with memory, which are ubiquitous in science and technology.

Efficient solution of Boolean satisfiability problems with digital memcomputing.

Boolean satisfiability is a propositional logic problem of interest in multiple fields, e.g., physics, mathematics, and computer science. Beyond a field of research, instances of the SAT problem, as it is known, require efficient solution methods in a variety of applications. It is the decision problem of determining whether a Boolean formula has a satisfying assignment, believed to require exponentially growing time for an algorithm to solve for the worst-case instances. Yet, the efficient solution of many classes of Boolean formulae eludes even the most successful algorithms, not only for the worst-case scenarios, but also for typical-case instances. Here, we introduce a memory-assisted physical system (a digital memcomputing machine) that, when its non-linear ordinary differential equations are integrated numerically, shows evidence for polynomially-bounded scalability while solving "hard" planted-solution instances of SAT, known to require exponential time to solve in the typical case for both complete and incomplete algorithms. Furthermore, we analytically demonstrate that the physical system can efficiently solve the SAT problem in continuous time, without the need to introduce chaos or an exponentially growing energy. The efficiency of the simulations is related to the collective dynamical properties of the original physical system that persist in the numerical integration to robustly guide the solution search even in the presence of numerical errors. We anticipate our results to broaden research directions in physics-inspired computing paradigms ranging from theory to application, from simulation to hardware implementation.

On the validity of memristor modeling in the neural network literature.

An analysis of the literature shows that there are two types of non-memristive models that have been widely used in the modeling of so-called "memristive" neural networks. Here, we demonstrate that such models have nothing in common with the concept of memristive elements: they describe either non-linear resistors or certain bi-state systems, which all are devices without memory. Therefore, the results presented in a significant number of publications are at least questionable, if not completely irrelevant to the actual field of memristive neural networks.

Stress-Testing Memcomputing on Hard Combinatorial Optimization Problems.

Memcomputing is a novel computing paradigm that employs time non-local dynamical systems to compute with and in memory. The digital version of these machines [digital memcomputing machines or (DMMs)] is scalable, and is particularly suited to solve combinatorial optimization problems. One of its possible realizations is by means of standard electronic circuits, with and without memory. Since these elements are non-quantum, they can be described by ordinary differential equations. Therefore, the circuit representation of DMMs can also be simulated efficiently on our traditional computers. We have indeed previously shown that these simulations only require time and memory resources that scale linearly with the problem size when applied to finding a good approximation to the optimum of hard instances of the maximum-satisfiability problem. The state-of-the-art algorithms, instead, require exponential resources for the same instances. However, in that work, we did not push the simulations to the limit of the processor used. Since linear scalability at smaller problem sizes cannot guarantee linear scalability at much larger sizes, we have extended these results in a stress-test up to 64×106 variables (corresponding to about 1 billion literals), namely the largest case that we could fit on a single core of an Intel Xeon E5-2860 with 128 GB of dynamic random-access memory (DRAM). For this test, we have employed a commercial simulator, Falcon of MemComputing, Inc. We find that the simulations of DMMs still scale linearly in both time and memory up to these very large problem sizes versus the exponential requirements of the state-of-the-art solvers. These results further reinforce the advantages of the physics-based memcomputing approach compared with traditional ones.

Taming a nonconvex landscape with dynamical long-range order: Memcomputing Ising benchmarks.

Recent work on quantum annealing has emphasized the role of collective behavior in solving optimization problems. By enabling transitions of clusters of variables, such solvers are able to navigate their state space and locate solutions more efficiently despite having only local connections between elements. However, collective behavior is not exclusive to quantum annealers, and classical solvers that display collective dynamics should also possess an advantage in navigating a nonconvex landscape. Here we give evidence that a benchmark derived from quantum annealing studies is solvable in polynomial time using digital memcomputing machines, which utilize a collection of dynamical components with memory to represent the structure of the underlying optimization problem. To illustrate the role of memory and clarify the structure of these solvers we propose a simple model of these machines that demonstrates the emergence of long-range order. This model, when applied to finding the ground state of the Ising frustrated-loop benchmarks, undergoes a transient phase of avalanches which can span the entire lattice and demonstrates a connection between long-range behavior and their probability of success. These results establish the advantages of computational approaches based on collective dynamics of continuous dynamical systems.

On the Universality of Memcomputing Machines.

Universal memcomputing machines (UMMs) represent a novel computational model in which memory (time nonlocality) accomplishes both tasks of storing and processing of information. UMMs have been shown to be Turing-complete, namely, they can simulate any Turing machine. In this paper, we first introduce a novel set theory approach to compare different computational models and use it to recover the previous results on Turing-completeness of UMMs. We then relate UMMs directly to liquid-state machines (or "reservoir-computing") and quantum machines ("quantum computing"). We show that UMMs can simulate both types of machines, hence they are both "liquid-" or "reservoir-complete" and "quantum-complete." Of course, these statements pertain only to the type of problems these machines can solve and not to the amount of resources required for such simulations. Nonetheless, the set-theoretic method presented here provides a general framework which describes the relationship between any computational models.

Accelerating deep learning with memcomputing.

Restricted Boltzmann machines (RBMs) and their extensions, often called "deep-belief networks", are powerful neural networks that have found applications in the fields of machine learning and artificial intelligence. The standard way to train these models resorts to an iterative unsupervised procedure based on Gibbs sampling, called "contrastive divergence", and additional supervised tuning via back-propagation. However, this procedure has been shown not to follow any gradient and can lead to suboptimal solutions. In this paper, we show an efficient alternative to contrastive divergence by means of simulations of digital memcomputing machines (DMMs) that compute the gradient of the log-likelihood involved in unsupervised training. We test our approach on pattern recognition using a modified version of the MNIST data set of hand-written numbers. DMMs sample effectively the vast phase space defined by the probability distribution of RBMs, and provide a good approximation close to the optimum. This efficient search significantly reduces the number of generative pretraining iterations necessary to achieve a given level of accuracy in the MNIST data set, as well as a total performance gain over the traditional approaches. In fact, the acceleration of the pretraining achieved by simulating DMMs is comparable to, in number of iterations, the recently reported hardware application of the quantum annealing method on the same network and data set. Notably, however, DMMs perform far better than the reported quantum annealing results in terms of quality of the training. Finally, we also compare our method to recent advances in supervised training, like batch-normalization and rectifiers, that seem to reduce the advantage of pretraining. We find that the memcomputing method still maintains a quality advantage (>1% in accuracy, corresponding to a 20% reduction in error rate) over these approaches, despite the network pretrained with memcomputing defines a more non-convex landscape using sigmoidal activation functions without batch-normalization. Our approach is agnostic about the connectivity of the network. Therefore, it can be extended to train full Boltzmann machines, and even deep networks at once.

An energy-resolved atomic scanning probe.

We propose a method to probe the local density of states (LDOS) of atomic systems that provides both spatial and energy resolution. The method combines atomic and tunneling techniques to supply a simple, yet quantitative and operational, definition of the LDOS for both interacting and non-interacting systems: It is the rate at which particles can be siphoned from the system of interest by a narrow energy band of non-interacting states contacted locally to the many-body system of interest. Ultracold atoms in optical lattices are a natural platform for implementing this broad concept to visualize the energy and spatial dependence of the atom density in interacting, inhomogeneous lattices. This includes models of strongly correlated condensed matter systems, as well as ones with non-trivial topologies.

Memcomputing Numerical Inversion With Self-Organizing Logic Gates.

We propose to use digital memcomputing machines (DMMs), implemented with self-organizing logic gates (SOLGs), to solve the problem of numerical inversion. Starting from fixed-point scalar inversion, we describe the generalization to solving linear systems and matrix inversion. This method, when realized in hardware, will output the result in only one computational step. As an example, we perform simulations of the scalar case using a 5-bit logic circuit made of SOLGs, and show that the circuit successfully performs the inversion. Our method can be extended efficiently to any level of precision, since we prove that producing -bit precision in the output requires extending the circuit by at most bits. This type of numerical inversion can be implemented by DMM units in hardware; it is scalable, and thus of great benefit to any real-time computing application.

Absence of periodic orbits in digital memcomputing machines with solutions.

In Traversa and Di Ventra [Chaos 27, 023107 (2017)] we argued, without proof, that if the non-linear dynamical systems with memory describing the class of digital memcomputing machines (DMMs) have equilibrium points, then no periodic orbits can emerge. In fact, the proof of such a statement is a simple corollary of a theorem already demonstrated in Traversa and Di Ventra [Chaos 27, 023107 (2017)]. Here, we point out how to derive such a conclusion. Incidentally, the same demonstration implies absence of chaos, a result we have already demonstrated in Di Ventra and Traversa [Phys. Lett. A 381, 3255 (2017)] using topology. These results, together with those in Traversa and Di Ventra [Chaos 27, 023107 (2017)], guarantee that if the Boolean problem the DMMs are designed to solve has a solution, the system will always find it, irrespective of the initial conditions.

Dehydration as a Universal Mechanism for Ion Selectivity in Graphene and Other Atomically Thin Pores.

Ion channels play a key role in regulating cell behavior and in electrical signaling. In these settings, polar and charged functional groups, as well as protein response, compensate for dehydration in an ion-dependent way, giving rise to the ion selective transport critical to the operation of cells. Dehydration, though, yields ion-dependent free-energy barriers and thus is predicted to give rise to selectivity by itself. However, these barriers are typically so large that they will suppress the ion currents to undetectable levels. Here, we establish that graphene displays a measurable dehydration-only mechanism for selectivity of K+ over Cl-. This fundamental mechanism, one that depends only on the geometry and hydration, is the starting point for selectivity for all channels and pores. Moreover, while we study selectivity of K+ over Cl- we find that dehydration-based selectivity functions for all ions, that is, cation over cation selectivity (e.g., K+ over Na+). Its likely detection in graphene pores resolves conflicting experimental results, as well as presents a new paradigm for characterizing the operation of ion channels and engineering molecular/ionic selectivity in filtration and other applications.

Solitonic Josephson-based meminductive systems.

Memristors, memcapacitors, and meminductors represent an innovative generation of circuit elements whose properties depend on the state and history of the system. The hysteretic behavior of one of their constituent variables, is their distinctive fingerprint. This feature endows them with the ability to store and process information on the same physical location, a property that is expected to benefit many applications ranging from unconventional computing to adaptive electronics to robotics. Therefore, it is important to find appropriate memory elements that combine a wide range of memory states, long memory retention times, and protection against unavoidable noise. Although several physical systems belong to the general class of memelements, few of them combine these important physical features in a single component. Here, we demonstrate theoretically a superconducting memory based on solitonic long Josephson junctions. Moreover, since solitons are at the core of its operation, this system provides an intrinsic topological protection against external perturbations. We show that the Josephson critical current behaves hysteretically as an external magnetic field is properly swept. Accordingly, long Josephson junctions can be used as multi-state memories, with a controllable number of available states, and in other emerging areas such as memcomputing, i.e., computing directly in/by the memory.

Polynomial-time solution of prime factorization and NP-complete problems with digital memcomputing machines.

We introduce a class of digital machines, we name Digital Memcomputing Machines, (DMMs) able to solve a wide range of problems including Non-deterministic Polynomial (NP) ones with polynomial resources (in time, space, and energy). An abstract DMM with this power must satisfy a set of compatible mathematical constraints underlying its practical realization. We prove this by making a connection with the dynamical systems theory. This leads us to a set of physical constraints for poly-resource resolvability. Once the mathematical requirements have been assessed, we propose a practical scheme to solve the above class of problems based on the novel concept of self-organizing logic gates and circuits (SOLCs). These are logic gates and circuits able to accept input signals from any terminal, without distinction between conventional input and output terminals. They can solve boolean problems by self-organizing into their solution. They can be fabricated either with circuit elements with memory (such as memristors) and/or standard MOS technology. Using tools of functional analysis, we prove mathematically the following constraints for the poly-resource resolvability: (i) SOLCs possess a global attractor; (ii) their only equilibrium points are the solutions of the problems to solve; (iii) the system converges exponentially fast to the solutions; (iv) the equilibrium convergence rate scales at most polynomially with input size. We finally provide arguments that periodic orbits and strange attractors cannot coexist with equilibria. As examples, we show how to solve the prime factorization and the search version of the NP-complete subset-sum problem. Since DMMs map integers into integers, they are robust against noise and hence scalable. We finally discuss the implications of the DMM realization through SOLCs to the NP = P question related to constraints of poly-resources resolvability.