Low Computational Cost Research Articles

Mean Field Initialization of the Annealed Importance Sampling Algorithm for an Efficient Evaluation of the Partition Function Using Restricted Boltzmann Machines

Probabilistic models in physics often require the evaluation of normalized Boltzmann factors, which in turn implies the computation of the partition function Z. Obtaining the exact value of Z, though, becomes a forbiddingly expensive task as the system size increases. A possible way to tackle this problem is to use the Annealed Importance Sampling (AIS) algorithm, which provides a tool to stochastically estimate the partition function of the system. The nature of AIS allows for an efficient and parallel implementation in Restricted Boltzmann Machines (RBMs). In this work, we evaluate the partition function of magnetic spin and spin-like systems mapped into RBMs using AIS. So far, the standard application of the AIS algorithm starts from the uniform probability distribution and uses a large number of Monte Carlo steps to obtain reliable estimations of Z following an annealing process. We show that both the quality of the estimation and the cost of the computation can be significantly improved by using a properly selected mean-field starting probability distribution. We perform a systematic analysis of AIS in both small- and large-sized problems, and compare the results to exact values in problems where these are known. As a result, we propose two successful strategies that work well in all the problems analyzed. We conclude that these are good starting points to estimate the partition function with AIS with a relatively low computational cost. The procedures presented are not linked to any learning process, and therefore do not require a priori knowledge of a training dataset.

Beyond CCSD(T) Accuracy at Lower Scaling with Auxiliary Field Quantum Monte Carlo.

We introduce a black-box auxiliary field quantum Monte Carlo (AFQMC) approach to perform highly accurate electronic structure calculations using configuration interaction singles and doubles (CISD) trial states. This method consistently provides more accurate energy estimates than coupled cluster singles and doubles with perturbative triples (CCSD(T)), often regarded as the gold standard in quantum chemistry. This level of precision is achieved at a lower asymptotic computational cost, scaling as O(N6) compared to the O(N7) scaling of CCSD(T). We provide numerical evidence supporting these findings through results for challenging main group and transition metal-containing molecules.

Nanoindentation and deformation behaviors of α$\alpha$‐Al2O3${\rm Al}_2 {\rm O}_3$: A molecular dynamic study

AbstractStudying the nanoindentation behavior of materials at the atomic level is crucial for advancing technologies involving energetic atom bombardment, ion implantation, and nanomechanical testing. Using molecular dynamics simulations with two typical rigid ion interatomic potentials (fixed charges), we showed that the mechanical responses and dislocation nucleation of alumina (‐) are highly temperature‐ and orientation‐dependent. The complex behavior in dislocation dynamics has been rationalized by computing the distribution of stacking fault energy. It is further found that SHIK potential provides better accuracy in describing dislocation dynamics at a much lower computational cost.

Autoencoder Reconstruction of Cosmological Microlensing Magnification Maps

Abstract Enhanced modeling of microlensing variations in light curves of strongly lensed quasars improves measurements of cosmological time delays, the Hubble Constant, and quasar structure. Traditional methods for modeling extragalactic microlensing rely on computationally expensive magnification map generation. With large data sets expected from wide-field surveys like the Vera C. Rubin Legacy Survey of Space and Time, including thousands of lensed quasars and hundreds of multiply imaged supernovae, faster approaches become essential. We introduce a deep-learning model that is trained on pre-computed magnification maps covering the parameter space on a grid of κ, γ, and s. Our autoencoder creates a low-dimensional latent space representation of these maps, enabling efficient map generation. Quantifying the performance of magnification map generation from a low dimensional space is an essential step in the roadmap to develop neural network-based models that can replace traditional feed-forward simulation at much lower computational costs. We develop metrics to study various aspects of the autoencoder generated maps and show that the reconstruction is reliable. Even though we observe a mild loss of resolution in the generated maps, we find this effect to be smaller than the smoothing effect of convolving the original map with a source of a plausible size for its accretion disk in the red end of the optical spectrum and larger wavelengths and particularly one suitable for studying the broad-line region of quasars. Used to generate large samples of on-demand magnification maps, our model can enable fast modeling of microlensing variability in lensed quasars and supernovae.

EPR Spectra Hyperfine Structure of Fluorine-Containing Nitroxide Radicals in Liquid Solutions: Prediction by DFT.

Spin trap method allows receiving important information about the structure of short-living intermediate radicals involved in chemical processes. Commonly, the product of a spin trap reaction with an intermediate radical is identified based on the hyperfine structure of its EPR spectrum. However, such identification can be significantly complicated for novel radicals whose spectra are unknown. In this work, we propose a semiquantitative low-cost computation method that allows predicting the hyperfine structure of EPR spectra of the α-phenyl-N-tert-butylnitrone (PBN) adducts with fluorine-containing radicals. The scheme was tested for several radicals containing from 0 to 4 fluorine atoms.

Simulating concrete cracking and failure under impact loading using a novel constitutive integration paradigm in non-ordinary state-based Peridynamics

Concrete cracking and failure under external loads are common in infrastructure, particularly under impact conditions where the variability of loads and the complexity of the cracking process present significant challenges. While the Bond-Based Peridynamic (BBPD) and Ordinary State-Based Peridynamic (OSBPD) models are widely used in concrete damage analysis due to their low computational cost, their simple constitutive laws limit their effectiveness in simulating concrete’s response to impact loading. To overcome this limitation, we propose a novel approach that integrates classical constitutive models into the Peridynamics (PD) framework. Specifically, we reformulate the Concrete Damage Plasticity Model 2 (CDPM2) within a non-ordinary state-based PD (NOSBPD) framework, resulting in the CDPM2-PD model. This model incorporates factors such as strain rate effects and strain hardening and introduces a bond damage criterion for both tensile and compressive damage. Through three numerical examples, the CDPM2-PD model demonstrates its ability to accurately capture the cracking and failure process of concrete under impact, showing strong agreement with experimental observations in areas such as failure mode transitions and crack patterns. These results validate the model’s effectiveness and offer a versatile method for integrating classical constitutive models into the NOSBPD framework for complex material analysis.

Equivalent Inclusion Method (EIM) for isotropic and anisotropic spatially oriented spheroidal inhomogeneities: A unified calculation module validated via comparisons to Finite Element (FE) simulations

This paper is centered on the Equivalent Inclusion Method due to Eshelby (Eshelby, 1957, 1959, 1961) to solve inhomogeneous ellipsoidal inclusion problems. The objective is to develop a unified EIM program for spatially oriented isotropic and anisotropic spheroids and validate each stage of the developments in a controlled and rigorous way. At first, the analytical calculations of spatial derivatives of the elliptic integrals, involved in the expression of the strain field in the isotropic infinite medium, are pushed as far as possible for an Oblate spheroid and coded as it was done by (Vincent et al., 2014) for the Prolate shape. Then, spatial orientation of the spheroid with respect to the global axis system attached to the infinite medium is introduced. Anisotropic metal inhomogeneities are finally dealt with, with the possibility to assign different crystallographic orientations. In this final configuration involving both the spatial orientation and anisotropy, three axis systems have to be managed simultaneously. Such a complexity legitimates a progressive evaluation towards this case. Thus, each new functionality introduced in the code is carefully validated by comparisons of the results to Finite Element reference solutions both inside and outside the inhomogeneity along different paths from the interface. This is done for an isotropic inhomogeneity without and with spatial orientation at first, and for an anisotropic inhomogeneity in the same way. These evaluations are presented for different shapes, aspect ratios, property contrasts. Such a complete evaluation involving at each stage various cases and examining the fields inside and outside the inhomogeneity constitutes an original contribution of the present work and allows to be confident in the proposed code (available on request). Another contribution of the paper is to analyze the influence of various factors (shape, aspect ratio, spatial orientation) on the fields distribution when both the anisotropy and the spatial orientation of the spheroidal inhomogeneity are combined. This last part illustrates also the efficiency of the EIM to deal with such analysis with both accuracy and a low calculation cost.

Three-Dimensional Object Recognition Using Orthogonal Polynomials: An Embedded Kernel Approach

Computer vision seeks to mimic the human visual system and plays an essential role in artificial intelligence. It is based on different signal reprocessing techniques; therefore, developing efficient techniques becomes essential to achieving fast and reliable processing. Various signal preprocessing operations have been used for computer vision, including smoothing techniques, signal analyzing, resizing, sharpening, and enhancement, to reduce reluctant falsifications, segmentation, and image feature improvement. For example, to reduce the noise in a disturbed signal, smoothing kernels can be effectively used. This is achievedby convolving the distributed signal with smoothing kernels. In addition, orthogonal moments (OMs) are a crucial technique in signal preprocessing, serving as key descriptors for signal analysis and recognition. OMs are obtained by the projection of orthogonal polynomials (OPs) onto the signal domain. However, when dealing with 3D signals, the traditional approach of convolving kernels with the signal and computing OMs beforehand significantly increases the computational cost of computer vision algorithms. To address this issue, this paper develops a novel mathematical model to embed the kernel directly into the OPs functions, seamlessly integrating these two processes into a more efficient and accurate approach. The proposed model allows the computation of OMs for smoothed versions of 3D signals directly, thereby reducing computational overhead. Extensive experiments conducted on 3D objects demonstrate that the proposed method outperforms traditional approaches across various metrics. The average recognition accuracy improves to 83.85% when the polynomial order is increased to 10. Experimental results show that the proposed method exhibits higher accuracy and lower computational costs compared to the benchmark methods in various conditions for a wide range of parameter values.

An AI-Driven Framework for Automated Human Violence Detection Using Advanced Deep Learning Model and IoT Systems

Human violence poses a significant threat to public safety, making its early detection critical in preventing harm and ensuring timely intervention. In today's world, it is a growing necessity that public safety can be achieved through intelligent surveillance systems. In this paper, we propose an AI driven framework on automated human violence detection through the combination of advanced deep learning techniques with IoT technologies. The system relies on YOLO (You Only Look Once) object detection model from processing live video feeds to determining violent actions like fights and assaults. A custom dataset, recorded and augmented by myself, was used to improve model reliability and performance. The framework was able to achieve 92% accuracy, proving its ability to produce real time results at low computational cost. A Telegram bot is used to transmit notifications and alerts instantly, boosting the level of security and intervention in time. The framework is designed for scalability and adaptability, being deployable in offices, schools or public areas. In the future, predictive analytics will be embedded, multi camera support will be added, and by using Cloud storage, system efficiency and scalability will further be enhanced. AI and IoT facilitates smart, responsive violence detection surveillance systems is a transformative piece of this research.

ChromoGen: Diffusion model predicts single-cell chromatin conformations.

Breakthroughs in high-throughput sequencing and microscopic imaging technologies have revealed that chromatin structures vary considerably between cells of the same type. However, a thorough characterization of this heterogeneity remains elusive due to the labor-intensive and time-consuming nature of these experiments. To address these challenges, we introduce ChromoGen, a generative model based on state-of-the-art artificial intelligence techniques that efficiently predicts three-dimensional, single-cell chromatin conformations de novo with both region and cell type specificity. These generated conformations accurately reproduce experimental results at both the single-cell and population levels. Moreover, ChromoGen successfully transfers to cell types excluded from the training data using just DNA sequence and widely available DNase-seq data, thus providing access to chromatin structures in myriad cell types. These achievements come at a remarkably low computational cost. Therefore, ChromoGen enables the systematic investigation of single-cell chromatin organization, its heterogeneity, and its relationship to sequencing data, all while remaining economical.

Evaluation of Semiempirical Quantum Mechanical Methods for Zr-Based Metal-Organic Framework Catalysts.

Zr-based metal-organic frameworks (MOFs) are typically employed in heterogeneous catalysis due to their porosity, chemical and thermal stability, and well-defined active sites. Density functional theory (DFT) is the workhorse to compute their electronic structure; however, it becomes very costly when dealing with reaction mechanisms involving large unit cells and vast configurational spaces. Semiempirical quantum mechanical (SQM) methods appear as an alternative approach to simulate such chemical systems at low computational cost, but their feasibility to model catalysis with MOFs is still unexplored. Thus, here we present a benchmark study on UiO-66 to evaluate the performance of SQM methods (PM6, PM7, GFN1-xTB, GFN2-xTB) against hybrid DFT (M06). We evaluate defective nodes, ligand exchange reactions, barrier heights, and host-guest interactions with metal nanoclusters. Despite some caveats, GFN1-xTB on properly constrained models is the best SQM method across all studied properties. Under proper supervision, this protocol holds promise for application in exploratory high-throughput screenings of Zr-based MOF catalysts, subject to further refinement with more accurate methods.

Lessons learned on obtaining reliable dynamic properties for ionic liquids.

Ionic liquids are nowadays investigated with respect to their use as electrolytes for high-performance energy storage materials. In this study, we provide a tutorial on how to calculate dynamic properties such as self-diffusion coefficients, ionic conductivities, transference numbers, as well as ion pair and ion cage dynamics, that all play a role in judging the applicability of ionic liquids as electrolytes. For the case of the ionic liquid \ce{[C2C1Im][NTf2]}, we investigate the performance of different force fields. Amongst them are non-polarizable models employing unity charges, a charge-scaled version of a non-polarizable model, a polarizable model and another non-polarizable model with refined Lennard-Jones parameters. We also study the influence of the system size on the dynamic properties. While all studied force field models capture qualitatively correct trends, only the polarizable force field and the non-polarizable force field with refined Lennard-Jones parameters provide quantitative agreement to reference data, making the latter model very attractive for the reason of lower computational costs.

Fourier or Wavelet bases as counterpart self-attention in spikformer for efficient visual classification

Energy-efficient spikformer has been proposed by integrating the biologically plausible spiking neural network (SNN) and artificial transformer, whereby the spiking self-attention (SSA) is used to achieve both higher accuracy and lower computational cost. However, it seems that self-attention is not always necessary, especially in sparse spike-form calculation manners. In this article, we innovatively replace vanilla SSA (using dynamic bases calculating from Query and Key) with spike-form Fourier transform, wavelet transform, and their combinations (using fixed triangular or wavelets bases), based on a key hypothesis that both of them use a set of basis functions for information transformation. Hence, the Fourier-or-Wavelet-based spikformer (FWformer) is proposed and verified in visual classification tasks, including both static image and event-based video datasets. The FWformer can achieve comparable or even higher accuracies (0.4%–1.5%), higher running speed (9%–51% for training and 19%–70% for inference), reduced theoretical energy consumption (20%–25%), and reduced graphic processing unit (GPU) memory usage (4%–26%), compared to the standard spikformer. Our result indicates the continuous refinement of new transformers that are inspired either by biological discovery (spike-form), or information theory (Fourier or Wavelet transform), is promising.

Fast Recurrent Field Transforms for Optical Flow on Edge GPUs

Abstract Optical flow reflects motion information in videos, serving as a crucial visual cue. However, high computational demands make dense optical flow estimation challenging to deploy on mobile platforms. We address this by introducing EdgeFlow, an optical flow network designed explicitly for edge GPUs. To enable efficient feature extraction, we propose a novel Low Memory Access Cost Net (LMAC-Net). Furthermore, we design a Convolutional Gated Core (CGC) for optical flow update, capable of obtaining high-quality features at low computational cost. During training and inference, we decouple the upsampling procedure. Inference speed is increased by only upsampling the final update outputs, reducing unnecessary computations. Our model demonstrates significant efficiency, utilizing only 52% of the parameters required by the baseline model. In generalization tests conducted on the Sintel dataset, EdgeFlow outperforms the baseline by achieving an 13% reduction in error rate. Additionally, the processing speed of EdgeFlow deployed on the Jetson Xavier NX platform is 7.5 times faster than the baseline. Finally, after using TensorRT to accelerate the model (with 3 recurrent updates), the inference speed is 98.67 FPS at 640×480 resolution. Source code is available at https://github.com/sklibra/EdgeFlow/tree/master.

PM6-ML: The Synergy of Semiempirical Quantum Chemistry and Machine Learning Transformed into a Practical Computational Method.

Machine learning (ML) methods offer a promising route to the construction of universal molecular potentials with high accuracy and low computational cost. It is becoming evident that integrating physical principles into these models, or utilizing them in a Δ-ML scheme, significantly enhances their robustness and transferability. This paper introduces PM6-ML, a Δ-ML method that synergizes the semiempirical quantum-mechanical (SQM) method PM6 with a state-of-the-art ML potential applied as a universal correction. The method demonstrates superior performance over standalone SQM and ML approaches and covers a broader chemical space than its predecessors. It is scalable to systems with thousands of atoms, which makes it applicable to large biomolecular systems. Extensive benchmarking confirms PM6-ML's accuracy and robustness. Its practical application is facilitated by a direct interface to MOPAC. The code and parameters are available at https://github.com/Honza-R/mopac-ml.

Enhanced soft Monte Carlo simulation coupled with support vector regression for structural reliability analysis

In structural reliability analysis, it is a major challenge to develop a general method that can ensure high computational accuracy and low computational cost for low failure probability and high-dimensional problems. In this study, a novel enhanced simulation method termed enhanced soft Monte Carlo simulation coupled with support vector regression (EMCS-SVR) is proposed. First, a generalized enhanced simulation (ES) scaling formula is proposed as an improved scheme. Furthermore, the soft Monte Carlo simulation is combined with generalized ES scaling formula and SVR model for evaluating the failure probability. The efficiency and accuracy of the ESMCS-SVR are verified by comparing with an existing popular method in four numerical examples and three engineering examples.

Optimal design of experiments in the context of machine-learning inter-atomic potentials: improving the efficiency and transferability of kernel based methods

Abstract Data-driven machine learning (ML) models of atomistic interactions are often based on flexible and non-physical functions that can relate nuanced aspects of atomic arrangements to predictions of energies and forces. As a result, these potentials are only as good as the training data (usually the results of so-called ab initio simulations), and we need to ensure that we have enough information to make a model sufficiently accurate, reliable and transferable. The main challenge stems from the fact that descriptors of chemical environments are often sparse, high-dimensional objects without a well-defined continuous metric. Therefore, it is rather unlikely that any ad hoc method for selecting training examples will be indiscriminate, and it is easy to fall into the trap of confirmation bias, where the same narrow and biased sampling is used to generate training and test sets. We will show that an approach derived from classical concepts of statistical planning of experiments and optimal design can help to mitigate such problems at a relatively low computational cost. The key feature of the method we will investigate is that it allows us to assess the quality of the data without obtaining reference energies and forces—a so-called offline approach. In other words, we are focusing on an approach that is easy to implement and does not require sophisticated frameworks that involve automated access to high performance computing.

Integration of Deep Learning Vision Systems in Collaborative Robotics for Real-Time Applications

Collaborative robotics and computer vision systems are increasingly important in automating complex industrial tasks with greater safety and productivity. This work presents an integrated vision system powered by a trained neural network and coupled with a collaborative robot for real-time sorting and quality inspection in a food product conveyor process. Multiple object detection models were trained on custom datasets using advanced augmentation techniques to optimize performance. The proposed system achieved a detection and classification accuracy of 98%, successfully processing more than 600 items with high efficiency and low computational cost. Unlike conventional solutions that rely on ROS (Robot Operating System), this implementation used a Windows-based Python framework for greater accessibility and industrial compatibility. The results demonstrated the reliability and industrial applicability of the solution, offering a scalable and accurate methodology that can be adapted to various industrial applications.

A three-stage plasma model based on one-way coupling of plasma dynamics, ionic motion, and fluid flow: Application to DBD plasma actuators

The scope of this paper is to present a comprehensive approach for simulating low-temperature atmospheric dielectric barrier discharge plasmas. The proposed methodology categorizes the primary physical phenomena: (i) discharge dynamics, (ii) ionic motion, and (iii) fluid flow, according to their respective time scales and simulates each independently. This allows for the use of distinct solution procedures tailored to each of the three stages of the problem. Such separation offers significant flexibility in choosing appropriate models and numerical schemes for each stage, enabling the simulation of complex geometries and large-scale applications without the excessive computational costs associated with a monolithic approach. As a case study, we apply the proposed algorithm to the surface dielectric barrier discharge plasma actuator for flow control, which is powered by alternating high voltages. The algorithm successfully described the actuator’s behavior while maintaining low computational cost. Additionally, a parametric study is conducted to examine the effect of key input parameters on the generated electrohydrodynamic force and the resulting velocity. Finally, an overall assessment of the three-stage model is provided, highlighting its efficiency and accuracy.

Sum-Rate Maximization in Massive MIMO with Joint Antenna Selection and Power Allocation

Telecommunications improvements in the last decades have led to an increase in the available data rate and reliability, and to reduce latency. Massive MIMO (mMIMO) has emerged as a promising solution to replace simple antenna setups. This arrangement involves a large number of transmitter antennas compared to the served terminals in the covered area, enabling higher data rates through the key characteristics of mMIMO. Despite its advantages, the utilization of such large arrays is energy-intensive. With fewer antennas, it is possible to achieve high data rates with reduced power consumption. Antenna selection is crucial for energy efficiency while maintaining computational efficiency. Additionally, power allocation in the downlink for each served terminal is an important step to consider. This work aims to propose a joint algorithm for selecting antennas and distributing energy among terminals, simplifying the process through low-complexity algorithms. The results show the penalty in sum-rate capacity for each proposed method. The two proposed algorithms demonstrate an adequate approach to the target value asymptotically as the number of transmitter antennas increases. This implies the possibility of achieving the joint process with significantly lower computational costs compared to perform antenna selection and power allocation separately.