Random Wire Research Articles

Exploring Randomly Wired Neural Networks for Climate Model Emulation

Abstract Exploring the climate impacts of various anthropogenic emissions scenarios is key to making informed decisions for climate change mitigation and adaptation. State-of-the-art Earth system models can provide detailed insight into these impacts but have a large associated computational cost on a per-scenario basis. This large computational burden has driven recent interest in developing cheap machine learning models for the task of climate model emulation. In this paper, we explore the efficacy of randomly wired neural networks for this task. We describe how they can be constructed and compare them with their standard feedforward counterparts using the ClimateBench dataset. Specifically, we replace the serially connected dense layers in multilayer perceptrons, convolutional neural networks, and convolutional long short-term memory networks with randomly wired dense layers and assess the impact on model performance for models with 1 million and 10 million parameters. We find that models with less-complex architectures see the greatest performance improvement with the addition of random wiring (up to 30.4% for multilayer perceptrons). Furthermore, of 24 different model architecture, parameter count, and prediction task combinations, only one had a statistically significant performance deficit in randomly wired networks relative to their standard counterparts, with 14 cases showing statistically significant improvement. We also find no significant difference in prediction speed between networks with standard feedforward dense layers and those with randomly wired layers. These findings indicate that randomly wired neural networks may be suitable direct replacements for traditional dense layers in many standard models. Significance Statement Modeling various greenhouse gas and aerosol emissions scenarios is important for both understanding climate change and making informed political and economic decisions. However, accomplishing this with large Earth system models is a complex and computationally expensive task. As such, data-driven machine learning models have risen in prevalence as cheap emulators of Earth system models. In this work, we explore a special type of machine learning model called randomly wired neural networks and find that they perform competitively for the task of climate model emulation. This indicates that future machine learning models for emulation may significantly benefit from using randomly wired neural networks as opposed to their more-standard counterparts.

Joint representations of color and form in mouse visual cortex described by random pooling from rods and cones.

Spatial transitions in color can aid any visual perception task, and its neural representation, the "integration of color and form," is thought to begin at primary visual cortex (V1). Integration of color and form is untested in mouse V1, yet studies show that the ventral retina provides the necessary substrate from green-sensitive rods and ultraviolet-sensitive cones. Here, we used two-photon imaging in V1 to measure spatial frequency (SF) tuning along four axes of rod and cone contrast space, including luminance and color. We first reveal that V1's sensitivity to color is similar to luminance, yet average SF tuning is significantly shifted lowpass for color. Next, guided by linear models, we used SF tuning along all four color axes to estimate the proportion of neurons that fall into classic models of color opponency, i.e., "single-," "double-," and "non-opponent." Few neurons (∼6%) fit the criteria for double opponency, which are uniquely tuned for chromatic borders. Most of the population can be described as a unimodal distribution ranging from strongly single-opponent to non-opponent. Consistent with recent studies of the rodent and primate retina, our V1 data are well-described by a simple model in which ON and OFF channels to V1 sample the photoreceptor mosaic randomly. Finally, an analysis comparing color opponency to preferred orientation and retinotopy further validates rods, and not cone M-opsin, as opponent with cone S-opsin in the upper visual field.NEW & NOTEWORTHY This study is the first to show that mouse V1 is highly sensitive to UV-green color contrast. Furthermore, it provides a detailed characterization of "color opponency," which is the putative neural basis for color perception. Finally, using an extremely simple yet novel random wiring model, we account for our observations.

Reduced Modeling for Electromagnetic Coupling to Randomly Wiring Automotive Cable Harness

The random wiring of automotive cable harness makes electromagnetic compatibility (EMC) analysis of the whole vehicle very complicated; thus, this paper proposes a simplification modeling technique to model the electromagnetic (EM) illumination on automotive cable harness. First, the stochastic process theory is applied to determine the cable route in a randomly bundling way. Then, the inductance and capacitance parameters of the cable harness at different location are established according to multiconductor transmission lines network (MTLN) theory. On this basis, the simplification modeling technique is developed to generate the electrical and geometrical parameters of the equivalent cable model. Finally, a model of shielded nine-conductor with randomly twisted and a model of unshielded nine-conductor with stochastic wiring in the vehicle are performed to validate the proposed approach by full-wave simulation. The presented method simplifies the complexity of modeling the complete cable harness significantly with a good accuracy.

Degeneracy measures in biologically plausible random Boolean networks

BackgroundDegeneracy—the ability of structurally different elements to perform similar functions—is a property of many biological systems. Highly degenerate systems show resilience to perturbations and damage because the system can compensate for compromised function due to reconfiguration of the underlying network dynamics. Degeneracy thus suggests how biological systems can thrive despite changes to internal and external demands. Although degeneracy is a feature of network topologies and seems to be implicated in a wide variety of biological processes, research on degeneracy in biological networks is mostly limited to weighted networks. In this study, we test an information theoretic definition of degeneracy on random Boolean networks, frequently used to model gene regulatory networks. Random Boolean networks are discrete dynamical systems with binary connectivity and thus, these networks are well-suited for tracing information flow and the causal effects. By generating networks with random binary wiring diagrams, we test the effects of systematic lesioning of connections and perturbations of the network nodes on the degeneracy measure.ResultsOur analysis shows that degeneracy, on average, is the highest in networks in which ~ 20% of the connections are lesioned while 50% of the nodes are perturbed. Moreover, our results for the networks with no lesions and the fully-lesioned networks are comparable to the degeneracy measures from weighted networks, thus we show that the degeneracy measure is applicable to different networks.ConclusionsSuch a generalized applicability implies that degeneracy measures may be a useful tool for investigating a wide range of biological networks and, therefore, can be used to make predictions about the variety of systems’ ability to recover function.

Modeling and statistical analysis of distribution parameters of random cable bundles based on image recognition technology

In this paper, a modeling method of actual bending random bundled wire harness is proposed. Firstly, based on image recognition technology, the three-dimensional coordinates of bending wire harness axis are reconstructed by using two photos of actual wire harness in side view and top view; then the random bundled wire harness is realized based on random transfer path method. Based on this modeling method, this paper analyzes the statistical characteristics of distribution parameters of bending random wire harness by Monte Carlo simulation, and finds that the variation trend of self inductance, mutual inductance and mutual capacitance along the line is consistent with the variation trend of wire harness height, while the trend of self capacitance is opposite; the coefficient of variation of self capacitance, self inductance and mutual inductance has negative correlation with wire harness height; the bundling randomness is not obvious It will change the mean value of self inductance and self capacitance, but reduce the mean value of mutual capacitance and mutual inductance.

A novel silicon interposer based high security integration approach for microsystem

Security problems of microsystem have gained widespread attention for its applications in many fields. However, current integration technologies cannot provide a security solution against physical invasive attack and side channel attack effectively. In this paper, a novel silicon interposer structure with embedded trenches was proposed to achieve high security microsystem integration. Chips were embedded in the trenches with metal shielding wires protected in all directions, which formed a Faraday cage. Experiment indicated that the Faraday cage could shield the electromagnetic radiation of the chips dramatically and the electromagnetic based side channel attack could be mitigated. Besides, metal shielding wires also formed a loop. Physical invasive attack on the chips would unavoidably damage the loop and could be detected. Since fine pitch random Hamiltonian metal wires could be fabricated on the silicon interposer, the designed structure had a high detection resolution.

Steady State and 2D Thermal Equivalence Circuit for Winding Heads—A New Modelling Approach

The study concerns the winding head thermal design of electrical machines in difficult thermal environments. The new approach is adapted for all basic shapes and solves the thermal behaviour of a random wire layout. The model uses the nodal method but does not use the common homogenization method for the winding slot. The layout impact can be precisely studied to find different hotspots. To achieve this a Delaunay triangulation provides the thermal links between adjoining wires in the slot. Voronoï tessellation gives a cutting to estimate thermal conductance between adjoining wires. This thermal behaviour is simulated in cell cutting and it is simplified with the thermal bridge notion to obtain a simple solving of these thermal conductances. The boundaries are imposed on the slot borders with Dirichlet condition. Then solving with many Dirichlet conditions is described. Some results show different possible applications with rectangular and round shapes, one ore many boundaries, different limit condition values and different layouts. The model can be integrated into a larger model that represents the stator to have best results.

Flexible neural connectivity under constraints on total connection strength.

Neural computation is determined by neurons' dynamics and circuit connectivity. Uncertain and dynamic environments may require neural hardware to adapt to different computational tasks, each requiring different connectivity configurations. At the same time, connectivity is subject to a variety of constraints, placing limits on the possible computations a given neural circuit can perform. Here we examine the hypothesis that the organization of neural circuitry favors computational flexibility: that it makes many computational solutions available, given physiological constraints. From this hypothesis, we develop models of connectivity degree distributions based on constraints on a neuron's total synaptic weight. To test these models, we examine reconstructions of the mushroom bodies from the first instar larva and adult Drosophila melanogaster. We perform a Bayesian model comparison for two constraint models and a random wiring null model. Overall, we find that flexibility under a homeostatically fixed total synaptic weight describes Kenyon cell connectivity better than other models, suggesting a principle shaping the apparently random structure of Kenyon cell wiring. Furthermore, we find evidence that larval Kenyon cells are more flexible earlier in development, suggesting a mechanism whereby neural circuits begin as flexible systems that develop into specialized computational circuits.

One-dimensional, surface emitting, disordered Terahertz lasers

Quantum cascade lasers are, by far, the most compact, powerful, and spectrally pure sources of radiation at terahertz frequencies, and, as such, they are of crucial importance for applications in metrology, spectroscopy, imaging, and astronomy, among many others. However, for many of those applications, particularly imaging, tomography, and near-field microscopy, undesired artifacts, resulting from the use of a coherent radiation source, can be detrimental. Random lasers can offer a concrete technological solution to the above issue. They, indeed, maintain a high degree of temporal coherence, as traditional lasers, while only exhibiting low spatial coherence, which can allow for the prevention of coherent artifacts, such as speckles. In this study, we report on the development of one-dimensional THz-frequency random wire lasers, patterned on the top surface of a double-metal quantum cascade laser with fully randomly arranged apertures, not arising from the perturbation of a regular photonic structure. By performing finite element method simulations, we engineer photonic patterns supporting strongly localized random modes in the 3.05–3.5 THz range. Multimode laser emission over a tunable-by-design band of about 400 GHz and with ∼2 mW of peak power has been achieved, associated with 10° divergent optical beam patterns. The achieved performances were then compared with those of perturbed Fabry–Perot disordered lasers, showing continuous-wave operation in the 3.5–3.8 THz range with an order of magnitude larger average power output than their random counterpart, and an irregular far field emission profile.

Characteristics of Electromagnetic Scattering from Vegetation Models using Random Wire Structures with Applications to Land Imaging SAR Systems

Clouds of both random curly wires and randomly-oriented straight wires with good conductivity are used to construct random volume models to simulate structures found in vegetation areas on the ground surface like naturally cultivated plants involving grass, trees, primary crops and chaff clouds. A variety of such random volumes are then subjected to incident electromagnetic plane wave of specific polarization to study the properties of electromagnetic scattering from such natural objects existing on the earth surface. It is shown that the frequency dependence of the Radar Cross Section (RCS) of such random structures has maxima around the frequencies corresponding to the natural modes of the wire structures constituting the volumetric model. On the other hand, the frequency behaviour of the RCS of these random clouds exhibits sharp peaks or anti-peaks over very narrow intervals of the frequency due to the generation of internal resonant modes between the finite-length pairs of wires constituting the random clouds. It is shown that such maxima, sharp peaks and anti-peaks of frequency behaviour of the backscattered field are very important to extract useful information for construction of microwave images and, also, for the classification of the vegetation areas that appear in earth remote sensing Synthetic Aperture Radar (SAR) images taken for the ground surface. The polarization properties of the fully polarimetric backscattering coefficients collected by a land imaging SAR system are studied through electromagnetic simulation.

Contact Wire Irregularity Stochastics and Effect on High-speed Railway Pantograph-Catenary Interactions

In high-speed rail operations, contact wire irregularity (CWI) in a catenary is a common disturbance to the pantograph–catenary interaction performance, which directly affects the quality of current collection. To describe the pointwise stochastics of CWI, the power spectral density (PSD) function for CWI is proposed, and its effect on the pantograph–catenary interaction is investigated. First, a preprocessing procedure is proposed to eliminate the redundant information in the measured irregularities based on the ensemble empirical mode decomposition (EEMD). Then, the upper envelope of the irregularity, which contains all the information regarding the dropper positions on the contact wire, is extracted. A form of the PSD function is suggested for contact wire irregularities. A methodology is proposed to include the effect of random irregularities in the assessment of the interaction performance of a pantograph–catenary. A developed target configuration under dead load (TCUD) method is used to calculate the initial configuration of the catenary, in which the dropper points on the contact wire are placed on their exact positions. Finally, the effect of the random contact wire irregularities on the contact force is investigated through 500 numerical simulations at each operating speed. The present results indicate that random irregularities have a direct impact on the pantograph–catenary contact, including an increment in the dispersion of the contact force statistics. The stochastic analysis shows that over 70% of the results with irregularities are worse than the ideal result without irregularities.

Loop Correlations in Random Wire Models

We introduce a family of loop soup models on the hypercubic lattice. The models involve links on the edges, and random pairings of the link endpoints on the sites. We conjecture that loop correlations of distant points are given by Poisson–Dirichlet correlations in dimensions three and higher. We prove that, in a specific random wire model that is related to the classical XY spin system, the probability that distant sites form an even partition is given by the Poisson–Dirichlet counterpart.

Randomly weighted receptor inputs can explain the large diversity of colour-coding neurons in the bee visual system

True colour vision requires comparing the responses of different spectral classes of photoreceptors. In insects, there is a wealth of data available on the physiology of photoreceptors and on colour-dependent behaviour, but less is known about the neural mechanisms that link the two. The available information in bees indicates a diversity of colour opponent neurons in the visual optic ganglia that significantly exceeds that known in humans and other primates. Here, we present a simple mathematical model for colour processing in the optic lobes of bees to explore how this diversity might arise. We found that the model can reproduce the physiological spectral tuning curves of the 22 neurons that have been described so far. Moreover, the distribution of the presynaptic weights in the model suggests that colour-coding neurons are likely to be wired up to the receptor inputs randomly. The perceptual distances in our random synaptic weight model are in agreement with behavioural observations. Our results support the idea that the insect nervous system might adopt partially random wiring of neurons for colour processing.

Anomalous transparency induced by cooperative disorders in phonon transport

Wave transport through disordered materials is of broad interest in science and engineering. It has been generally believed that long-range transport cannot exist in low-dimensional disordered systems due to the Anderson localization of all states. However, in this work we report the anomalous transparency of phonon transport induced by cooperative effects of mass and force-constant disorders in completely random harmonic wires. In contrast to the full localization in mass-disordered wire, an anomalous transparent phonon state can emerge due to the intrinsic cooperation of disordered mass and force constants. We obtain the frequency of the transparent state, tuned by the disordered force constants and masses. Consequently, the phonon transport in the completely disordered system falls into three different regimes: the full localization, resonant, and sub-ballistic regimes. While there exists a single nonzero transparent frequency in the resonant regime, transparent states in a wide range of frequencies can be present in the sub-ballistic regime, which significantly enhances the thermal conductivity of disordered wires. These findings may open new doors to modulate thermal transport and design thermal devices, including thermal filter and high-thermal-conductivity materials.

The Cellular Automaton Pulsing Model, Experiments with DDLab

The cellular automaton (CA) pulsing model (arXiv:1806.06416) described the surprising phenomenon of spontaneous, sustained and robust rhythmic oscillations, pulsing dynamics, when random wiring is applied to a 2D `glider' rule running in a 3-value totalistic CA. Case studies, pulsing measures, possible mechanisms, and implications for oscillatory networks in biology were presented. In this paper we summarise the results, extend the entropy-density and density-return map plots to include a linked history, look at totalistic glider rules with neighborhoods of 3, 4 and 5, as well as 6 and 7 studied previously, introduce methods to automatically recognise the wavelength, and extend results for randomly asynchronous updating. We show how the model is implemented in DDLab to validate results, output data, and allow experiments and research by others.

Statistical Characterization of Indian Residential Networks for Powerline Communication

Despite different powerline channel modeling techniques, developed so far, there are still specific dynamic, varying parameters (viz. the random load variation and inconsistent electrical wiring) to be studied for a valid and reliable power line communication (PLC) model. Statistical characterization of PLC channel may provide the required background for refinement of these existent models. In this paper, the Indian residential networks are statistically analyzed in the frequency range of 1-100 MHz. This also includes the comprehensive analysis of line impedance, stationary noise, channel capacity and average channel gain. From the measurements, the noise spectrum density is found to be less than -90 dBm at a frequency less than 1 MHz and is almost constant after 70 MHz. The minimum and maximum channel capacity of the network is 71.5 Mbps and 97.7 Mbps respectively. The Average channel gain is estimated at -30 dB. The paper also reviews the channel transfer function developed by top-down and bottom-top approaches. Finally, some additional factors influencing the PLC channel are also discussed.

Computational Investigation of the Morphology, Efficiency, and Properties of Silver Nano Wires Networks in Transparent Conductive Film

Random networks of silver nano wires have been considered for use in transparent conductive films as an alternative to Indium Tin Oxide (ITO), which is unsuitable for flexible devices. However, the random distribution of nano wires makes such conductive films non-uniform. As electrical conductivity is achieved through a percolation process, understanding the scale-dependency of the macroscopic properties (like electrical conductivity) and the exact efficiency of the network (the proportion of nano wires that participate in electrical conduction) is essential for optimizing the design. In this paper, we propose a computational method for identifying the representative volume element (RVE) of nano wire networks. This defines the minimum pixel size in devices using such transparent electrodes. The RVE is used to compute the macroscopic properties of films and to quantify the electrically conducting efficiency of networks. Then, the sheet resistance and transparency of networks are calculated based on the predicted RVEs, in order to analyze the effects of nano wire networks on the electrical and optical properties of conductive films. The results presented in this paper provide insights that help optimizing random nano wire networks in transparent conductive films for achieving better efficiencies.

On Random Wiring in Practicable Folded Clos Networks for Modern Datacenters

Big scale, high performance and fault-tolerance, low-cost and graceful expandability are pursued features in current datacenter networks (DCN). Although there have been many proposals for DCNs, most modern installations are equipped with classical folded Clos networks. Recently, regular random topologies, as the Jellyfish, have been proposed for DCNs. However, their completely unstructured nature entails serious design problems. In this paper we propose Random Folded Clos (RFC) and Hydra networks in which the interconnection between certain switches levels is made randomly. Both RFCs and Hydras preserve important properties of Clos networks that provide a straightforward deadlock-free multi-path routing. The proposed networks leverage randomness to be gracefully expandable, thereby allowing for fine grain upgrading. RFCs and Hydras are compared in the paper, in topological and cost terms, against fat-trees, orthogonal fat-trees and random regular networks. Also, experiments are carried out to simulate their performance under synthetic traffic patterns emulating common loads present in warehouse scale computers. These theoretical and empirical studies reveal the interest of these topologies, concluding that Hydra constitutes a practicable alternative to current datacenter networks since it appropriately balance all the main design requirements. Moreover, Hydras perform better than the fat-trees, their natural competitor, being able to connect the same or more computing nodes with significant lower cost and latency while exhibiting comparable throughput.

Microscopic Evaluation of Electrical and Thermal Conduction in Random Metal Wire Networks.

Ideally, transparent heaters exhibit uniform temperature, fast response time, high achievable temperatures, low operating voltage, stability across a range of temperatures, and high optical transmittance. For metal network heaters, unlike for uniform thin-film heaters, all of these parameters are directly or indirectly related to the network geometry. In the past, at equilibrium, the temperature distributions within metal networks have primarily been studied using either a physical temperature probe or direct infrared (IR) thermography, but there are limits to the spatial resolution of these cameras and probes, and thus, only average regional temperatures have typically been measured. However, knowledge of local temperatures within the network with a very high spatial resolution is required for ensuring a safe and stable operation. Here, we examine the thermal properties of random metal network thin-film heaters fabricated from crack templates using high-resolution IR microscopy. Importantly, the heaters achieve predominantly uniform temperatures throughout the substrate despite the random crack network structure (e.g., unequal sized polygons created by metal wires), but the temperatures of the wires in the network are observed to be significantly higher than the substrate because of the significant thermal contact resistance at the interface between the metal and the substrate. Last, the electrical breakdown mechanisms within the network are examined through transient IR imaging. In addition to experimental measurements of temperatures, an analytical model of the thermal properties of the network is developed in terms of geometrical parameters and material properties, providing insights into key design rules for such transparent heaters. Beyond this work, the methods and the understanding developed here extend to other network-based heaters and conducting films, including those that are not transparent.

Modeling field-to-wire coupling in random bundles of wires

A simple and computationally efficient method to represent the geometry of hand-made random wire bundles is described, which can be exploited within a multiconductor-transmission-line model for prediction of the currents induced in bundle terminal units by an external electromagnetic field. As an illustrative example, a random bundle running above a metallic ground and illuminated by a uniform plane wave is considered. By processing the data resulting from repeated simulations (Monte Carlo approach), some significant properties of the statistics of the induced currents are observed. Particularly, it is shown that in the so-called “well randomized” bundles the probability density function of the magnitude of the induced currents tends to be Normal at low frequency and Rayleigh at high frequency.