SPINBIS: Spintronics-Based Bayesian Inference System With Stochastic Computing

Bayesian inference is an effective approach for solving statistical learning problems especially with uncertainty and incompleteness. However, inference efficiencies are physically limited by the bottlenecks of conventional computing platforms. In this paper, an emerging Bayesian inference system is proposed by exploiting spintronics based stochastic computing. A stochastic bitstream generator is realized as the kernel components by leveraging the inherent randomness of spintronics devices. The proposed system is evaluated by typical applications of data fusion and Bayesian belief networks. Simulation results indicate that the proposed approach could achieve significant improvement on inference efficiencies in terms of power consumption and inference speed.

Digital film processing is characterized by a resolution of at least 2 K (2048 × 1536 pixels per frame at 30 bit/pixel and 24 pictures/s, data rate of 2.2 Gbit/s); higher resolutions of 4 K (8.8 Gbit/s) and even 8 K (35.2Gbit/s) are on their way. Real-time processing at this data rate is beyond the scope of today's standard and DSP processors, and ASICs are not economically viable due to the small market volume. Therefore, an FPGA-based approach was followed in the FlexFilm project. Different applications are supported on a single hardware platform by using different FPGA configurations. The multiboard, multi-FPGA hardware/software architecture, is based on Xilinx Virtex-II Pro FPGAs which contain the reconfigurable image stream processing data path, large SDRAM memories for multiple frame storage, and a PCI-Express communication backbone network. The FPGA-embedded CPU is used for control and less computation intensive tasks. This paper will focus on three key aspects: (a) the used design methodology which combines macro component configuration and macrolevel floorplaning with weak programmability using distributed microcoding, (b) the global communication framework with communication scheduling, and (c) the configurable multistream scheduling SDRAM controller with QoS support by access prioritization and traffic shaping. As an example, a complex noise reduction algorithm including a 2.5-dimension discrete wavelet transformation (DWT) and a full 16 × 16 motion estimation (ME) at 24 fps, requiring a total of 203 Gops/s net computing performance and a total of 28 Gbit/s DDR-SDRAM frame memory bandwidth, will be shown.

Low reverse-bias series resistance tunnel junctions (TJs) are the key to improving the performances of high efficiency multi-junction semiconductor laser diodes (MJLDs). In this paper, InGaAs QW TJ and InGaAs DQW TJ with single InGaAs layer and double InGaAs layers inserted into GaAs TJs separately, are proposed. TJ chips were fabricated by metal organic chemical vapor deposition (MOCVD) technology and semiconductor process. The measurement results of the devices display that the operating voltage of the InGaAs QW TJ and the InGaAs DQW TJ is lower than that of the GaAs TJ under the same injection current, whether it is a small current or a large current, and the InGaAs DQW TJ operating voltage is lower than that of the InGaAs QW TJ. Both GaAs TJ and InGaAs DQW TJ were applied to 1060 nm dual active region semiconductor laser diode. The ridge lasers with a strip width of 100μm and a cavity length of 2 mm were fabricated. The working voltage is reduced from 3.81 V to 3.38 V at 1 A drive current. Further experimental results indicate that the reverse-bias series resistance of InGaAs QW TJ and InGaAs DQW TJ is lower than that of GaAs TJ, and the performances of InGaAs DQW TJ are the best. This is of great significance to reduce the heat loss of MJLDs and improve its performances.

In many engineering problems, unknown parameters of a model are inferred in order to make predictions, to design controllers, or to optimize the model. When parameters are distributed (continuous) or very high-dimensional (discrete) and quantities of interest are low-dimensional, parameters need not be fully resolved to make accurate estimates of quantities of interest. In this work, we extend goal-oriented inference---the process of estimating predictions from observed data without resolving the parameter, previously justified theoretically in the linear setting---to Bayesian statistical inference problem formulations with nonlinear experimental and prediction processes. We propose to learn the joint density of data and predictions offline using Gaussian mixture models. When data are observed online, we condition the representation to arrive at a probabilistic description of predictions given observed data. Our approach enables real-time estimation of uncertainty in quantities of interest and renders tractable high-dimensional PDE-constrained Bayesian inference when there exist low-dimensional output quantities of interest. We demonstrate the method on a realistic problem in carbon capture and storage for which existing methods of Bayesian parameter estimation are intractable.

Efficient Bayesian inference using adversarial machine learning and low-complexity surrogate models

Conceptual development regarding single electron transfer phenomena is presented. It is shown that energy necessary to place a single electronic charge on one plate of a capacitor with equal opposite charge on its opposite plate is not a clearly distinguishable event at micro- and milliscales. But it becomes a meaningful event at the nanoscale due to the significant amount of energy involved. Further, it is shown that the existence of a voltage requirement for tunneling to occur across the plates of a capacitor, the so-called Coulomb blockade effect, is a noticeable phenomenon exclusive to nanoscale. Moving further, it is found that the Coulomb blockade is observable at or near room temperatures only in the scale of nano dimensions. From this understanding, the notion of a tunnel junction is put forward as a barrier in the form of an electrical potential or thin dielectric film across which tunneling occurs. The tunnel junction is modeled as an ideal capacitor with a parallel-connected tunnel resistance whose value must be ≫4.2 kΩ for Coulomb blockade to become recognizable. The capacitor of the tunnel junction behaves in a different way from a normal capacitor. Upon excitation by a constant current source , the voltage across this capacitor oscillates between two values, which are referred to as single electron tunneling oscillations . Applying the tunnel junction model, the operation of a quantum dot circuit consisting of a quantum dot and two tunnel junctions is analyzed. In both cases, for electron tunneling into the quantum dot across one tunnel junction and for electron tunneling off the quantum dot across the other tunnel junction, Coulomb blockade occurs in a quantum dot circuit like a nanocapacitor. The energy band diagram for a small quantum dot circuit exhibits a discretized nature. The electron tunneling does not take place in a continuous fashion but in discrete voltage steps. The resulting current–voltage characteristic of the quantum dot circuit has the shape of a staircase which is called the Coulomb staircase .

Bayesian Linear Regression analysis of the effect of silica fume on concrete creep predictions and structures’ safety

The olfactory information that is received by the insect brain is encoded in the form of spatiotemporal patterns in the projection neurons of the antennal lobe. These dense and overlapping patterns are transformed into a sparse code in Kenyon cells in the mushroom body. Although it is clear that this sparse code is the basis for rapid categorization of odors, it is yet unclear how the sparse code in Kenyon cells is computed and what information it represents. Here we show that this computation can be modeled by sequential firing rate patterns using Lotka-Volterra equations and Bayesian online inference. This new model can be understood as an ‘intelligent coincidence detector’, which robustly and dynamically encodes the presence of specific odor features. We found that the model is able to qualitatively reproduce experimentally observed activity in both the projection neurons and the Kenyon cells. In particular, the model explains mechanistically how sparse activity in the Kenyon cells arises from the dense code in the projection neurons. The odor classification performance of the model proved to be robust against noise and time jitter in the observed input sequences. As in recent experimental results, we found that recognition of an odor happened very early during stimulus presentation in the model. Critically, by using the model, we found surprising but simple computational explanations for several experimental phenomena.

We consider situations in Bayesian analysis where we have a family of priors $\nu_h$ on the parameter $\theta$, where $h$ varies continuously over a space $\mathcal{H}$, and we deal with two related problems. The first involves sensitivity analysis and is stated as follows. Suppose we fix a function $f$ of $\theta$. How do we efficiently estimate the posterior expectation of $f(\theta)$ simultaneously for all $h$ in $\mathcal{H}$? The second problem is how do we identify subsets of $\mathcal{H}$ which give rise to reasonable choices of $\nu_h$? We assume that we are able to generate Markov chain samples from the posterior for a finite number of the priors, and we develop a methodology, based on a combination of importance sampling and the use of control variates, for dealing with these two problems. The methodology applies very generally, and we show how it applies in particular to a commonly used model for variable selection in Bayesian linear regression, and give an illustration on the US crime data of Vandaele.

Efficient Bayesian inference with latent Hamiltonian neural networks in No-U-Turn Sampling

To improve the sampling efficiency of Markov Chain Monte Carlo in complex parameter spaces, this paper proposes an adaptive sampling method that integrates a swarm intelligence mechanism called the BeeSwarm-MH algorithm. The method combines global exploration by scout bees with local exploitation by worker bees. It employs multi-stage perturbation intensities and adaptive step-size tuning to enable efficient posterior sampling. Focusing on Bayesian inference for parameter estimation in the soliton solutions of the two-dimensional complex Ginzburg–Landau equation, we design a dedicated inference framework to systematically compare the performance of BeeSwarm-MH with the classical Metropolis–Hastings algorithm. Experimental results demonstrate that BeeSwarm-MH achieves comparable estimation accuracy while significantly reducing the required number of iterations and total computation time for convergence. Moreover, it exhibits superior global search capabilities and adaptive features, offering a practical approach for efficient Bayesian inference in complex physical models.

<div>We present a framework to solve geophysical inverse problems using the Hamiltonian Monte Carlo (HMC) method, with a focus on Bayesian tomography. Recent work in the geophysical community has shown the potential for gradient-based Monte Carlo sampling for a wide range of inverse problems across several fields.</div><div> </div><div>Many high-dimensional (non-linear) problems in geophysics have readily accessible gradient information which is unused in classical probabilistic inversions. Using HMC is a way to help improve traditional Monte Carlo sampling while increasing the scalability of inference problems, allowing access to uncertainty quantification for problems with many free parameters (>10'000). The result of HMC sampling is a collection of models representing the posterior probability density function, from which not only "best" models can be inferred, but also uncertainties and potentially different plausible scenarios, all compatible with the observed data. However, the amount of tuning parameters required by HMC, as well as the complexity of existing statistical modeling software, has limited the geophysical community in widely adopting a specific tool for performing efficient large-scale Bayesian inference.</div><div> </div><div>This work attempts to make a step towards filling that gap by providing an HMC sampler tailored for geophysical inverse problems (by e.g. supplying relevant priors and visualizations) combined with a set of different forward models, ranging from elastic and acoustic wave propagation to magnetic anomaly modeling, traveltimes, etc.. The framework is coded in the didactic but performant languages Julia and Python, with the possibility for the user to combine their own forward models, which are linked to the sampler routines by proper interfaces. In this way, we hope to illustrate the usefulness and potential of HMC in Bayesian inference. Tutorials featuring an array of physical experiments are written with the aim of both showcasing Bayesian inference and successful HMC usage. It additionally includes examples on how to speed up HMC e.g. with automated tuning techniques and GPU computations.</div>

Bayesian consideration of unknown sensor characteristics in fatigue-related structural health monitoring

Summary The inferential model (IM) framework provides valid prior-free probabilistic inference by focusing on predicting unobserved auxiliary variables. But, efficient IM-based inference can be challenging when the auxiliary variable is of higher dimension than the parameter. Here we show that features of the auxiliary variable are often fully observed and, in such cases, a simultaneous dimension reduction and information aggregation can be achieved by conditioning. This proposed conditioning strategy leads to efficient IM inference and casts new light on Fisher's notions of sufficiency, conditioning and also Bayesian inference. A differential-equation-driven selection of a conditional association is developed, and validity of the conditional IM is proved under some conditions. For problems that do not admit a conditional IM of the standard form, we propose a more flexible class of conditional IMs based on localization. Examples of local conditional IMs in a bivariate normal model and a normal variance components model are also given.

• Differentiable multi-fidelity inference: BMFIA enables gradient-based Bayesian inference for high-dimensional, non-differentiable multi-physics problems using low-fidelity model gradients. • Efficiency with minimal HF data: Accurate posterior estimates are achieved from only 100–300 HF–LF simulations with orders-of-magnitude speed-ups over HF-only inference. • Accuracy despite crude LF models: Robust posterior reconstruction is maintained even with weakly correlated or simplified low-fidelity models. High-dimensional Bayesian inverse analysis ( dim ≫ 100 ) is mostly unfeasible for computationally demanding, nonlinear physics-based high-fidelity (HF) models. Usually, the use of more efficient gradient-based inference schemes is impeded if the multi-physics models are provided by complex legacy codes. Adjoint-based derivatives are either exceedingly cumbersome to derive or nonexistent for practically relevant large-scale nonlinear and coupled multi-physics problems. Similarly, holistic automated differentiation w. r. t. primary variables of multi-physics codes is usually not yet an option and requires extensive code restructuring if not considered from the outset in the software design. This absence of differentiability further exacerbates the already present computational challenges. To overcome the existing limitations, we propose a novel inference approach called Bayesian multi-fidelity inverse analysis (BMFIA) , which leverages simpler and computationally cheaper lower-fidelity (LF) models that are designed to provide model derivatives. BMFIA learns a simple, probabilistic dependence of the LF and HF models, which is then employed in an altered likelihood formulation to statistically correct the inaccurate LF response. From a Bayesian viewpoint, this dependence represents a multi-fidelity (MF) conditional density (discriminative model). We demonstrate how this MF conditional density can be learned robustly in the small data regime from only a few HF and LF simulations (50 to 300), which would not be sufficient for naive surrogate approaches. The formulation is fully differentiable and allows the flexible design of a wide range of LF models. We demonstrate that BMFIA solves Bayesian inverse problems for scenarios that used to be prohibitive, such as finely-resolved and hence high-dimensional spatial reconstruction problems in two-dimensional Euclidean domains with static posteriors, given nonlinear and transient coupled poro-elastic media physics. We show that the resulting static MF posteriors are in excellent agreement with the (usually inaccessible) HF posteriors or ground-truth data and note that extending the framework to arbitrary three-dimensional domains is a natural and important direction for future work.