Fast Computation Research Articles

New clinical trial design borrowing information across patient subgroups based on fusion-penalized regression models.

In cancer research, basket trials aim to assess the efficacy of a drug using baskets, wherein patients are organized into subgroups according to their tumor type. In this context, using information borrowing strategy may increase the probability of detecting drug efficacy in active baskets, by shrinking together the estimates of the parameters characterizing the drug efficacy in baskets with similar drug activity. Here, we propose to use fusion-penalized logistic regression models to borrow information in the setting of a phase 2 single-arm basket trial with binary outcome. We describe our proposed strategy and assess its performance via a simulation study. We assessed the impact of heterogeneity in drug efficacy, prevalence of each tumor types and implementation of interim analyses on the operating characteristics of our proposed design. We compared our approach with two existing designs, relying on the specification of prior information in a Bayesian framework to borrow information across similar baskets. Notably, our approach performed well when the effect of the drug varied greatly across the baskets. Our approach offers several advantages, including limited implementation efforts and fast computation, which is essential when planning a new trial as such planning requires intensive simulation studies.

Comparison of Extremely Randomized Survival Trees and Random Survival Forests: A Simulation Study

Abstract. Purpose: This simulation study investigates the Extremely Randomized Survival Trees (EST) model, a machine learning technique expected to handle survival analysis, particularly in large survival datasets, effectively. The study compares the performance of the EST model with that of the Random Survival Forest (RSF) model, focusing on the C-index value to determine which model performs better. Methods: The analysis begins with the generation of 540 simulated datasets, created by combining three levels of sample sizes, two levels of censoring proportions, three types of hazard functions, and 30 repetitions for each scenario. The simulation data were split into 80% training and 20% testing data. The training data were used to build the EST and RSF models, while the test data were used to evaluate their performance. The model with the highest C-index value was deemed the best performer, as a higher C-index indicates superior model performance. Result: The results indicate that the sample size, type of hazard function, and the method used influence that model performance. The EST model significantly outperformed the RSF model when the sample size was large, though no significant difference was observed when the sample size was small or medium. Additionally, the EST model consistently demonstrated faster computation times across all simulation scenarios. Novelty: This study provides a pioneering exploration into applying decision tree algorithms, specifically EST and RSF, in survival analysis. While these methods have been extensively studied in regression and classification contexts, their application in survival analysis remains relatively unexplored.

Binning method for artifact-free time-tag based correlation function calculations.

Correlation functions are nowadays routinely computed using time-tagged photon information instead of a hardware autocorrelator. The algorithm developed by Laurence et al. [Opt. Lett.31, 829 (2006)10.1364/OL.31.000829] is a powerful example. Despite its ease of implementation and fast computation process, it presents a prevalent noisy feature at the short time-lag range when computed on commonly used logarithmically spaced bins. We identified that arbitral logarithmic spacing produces the mismatch between the edges of generated bins and acquisition frequency, resulting in an aliasing artifact at the short time-lag range of the correlation function. We introduce a binning method that considers the acquisition frequency during the bin generation. It effectively eliminates the artifact and improves the accuracy of the autocorrelation. Applying the binning method herein can be particularly crucial when one extracts photophysical processes from fluorescence correlation spectroscopy or the diffusion coefficient of nanoparticles from dynamic light scattering at the time range below 10-5 s lag time.

An Antiferromagnetic Neuromorphic Memory Based on Perpendicularly Magnetized CoO.

Antiferromagnets (AFMs) are ideal materials to boost neuromorphic computing toward the ultrahigh speed and ultracompact integration regime. However, developing a suitable AFM neuromorphic memory remains an aspirational but challenging goal. In this work, we construct such a memory based on the CoO/Pt heterostructure, in which the collinear insulating AFM CoO shows a strong perpendicular anisotropy facilitating its electrical readout and writing. Utilizing the unique nonlinear response and bipolar fading memory properties of the device, we demonstrate a multidimensional reservoir computing beyond the traditional binary paradigm. These results are expected to pave the way toward next-generation fast and massive neuromorphic computing.

Machine learning modeling for fuel cell-battery hybrid power system dynamics in a Toyota Mirai 2 vehicle under various drive cycles

Electrification is considered essential for the decarbonization of mobility sector, and understanding and modeling the complex behavior of modern fuel cell-battery electric-electric hybrid power systems is challenging, especially for product development and diagnostics requiring quick turnaround and fast computation. In this study, a novel modeling approach is developed, utilizing supervised machine learning algorithms, to replicate the dynamic characteristics of the fuel cell-battery hybrid power system in a 2021 Toyota Mirai 2nd generation (Mirai 2) vehicle under various drive cycles. The entire data for this study is collected by instrumenting the Mirai vehicle with in-house data acquisition devices and tapping into the Mirai controller area network bus during chassis dynamometer tests. A multi-input - multi-output, feed-forward artificial neural network architecture is designed to predict not only the fuel cell attributes, such as average minimum cell voltage, coolant and cathode air outlet temperatures, but also the battery hybrid system attributes, including lithium-ion battery pack voltage and temperature with the help of 15 system operating parameters. Over 21,0000 data points on various drive cycles having combinations of transient and near steady-state driving conditions are collected, out of which around 15,000 points are used for training the network and 6,000 for the evaluation of the model performance. Various data filtration techniques and neural network calibration processes are explored to condition the data and understand the impact on model performance. The calibrated neural network accurately predicts the hybrid power system dynamics with an R-squared value greater than 0.98, demonstrating the potential of machine learning algorithms for system development and diagnostics.

Complex dynamics in nonlinear small time-delayed optoelectronic oscillator and application in fast reservoir computing and pulse generation

We investigated a time-delayed optoelectronic oscillator (OEO) that displays a wide range of complex dynamic behavior under small time delay. The phase-space trajectory distributions in different dynamic regimes were compared which brings a new perspective on the underlying mechanism of the transition process. It was found that bifurcation is always possible no matter how small the time delay is even if the universal adiabatic approximation model is invalid. Hereby we proposed a versatile simple oscillator which has a potential capacity as memory carrier and high-dimensional state spatial mapping ability that brings 1000 times computing-efficiency improvements of reservoir computing over the large time delay one. Furthermore, we demonstrated a new approach for a tunable optoelectronic pulse generator (repetition rate at 0.2 MHz and 0.25 GHz) which depends critically on time-delayed input electrical pulse. The proposed oscillator is also a promising system for the applications of fast chaos-based communication.

Fast calculation of a cylindrical hologram by a preloaded stochastic gradient descent with skip connection

Cylindrical holography has received widespread attention due to its unique 360° viewing zone. To achieve commercial quality requirements, introducing stochastic gradient descent (SGD) is a potential approach for computer-generated cylindrical holography (CGCH). However, SGD applied to CGCH suffers from both slow convergence speed and unstable convergence, severely impacting its application. To address these issues, a preloaded SGD method with skip connection is proposed for fast calculation of cylindrical holograms in this paper. Preloaded-SGD (PSGD) exhibits a significant enhancement in convergence speed compared to the conventional SGD. Furthermore, the skip connection prevents oscillations from occurring by directly connecting the input and output, which is highly beneficial for obtaining high-quality holograms in the later stages of convergence. Numerical simulations demonstrate the effectiveness of our proposed method. PSGD with skip connection(SC-PSGD) achieves a 6.3-fold acceleration over conventional SGD. Notably, our proposed method has broad application prospects in cylindrical holographic displays and 3D displays.

Data safety prediction using YOLOv7+G3HN for traffic roads

Pulau Pinang has introduced several measures to enhance traffic safety and promote sustainability, including the installation of CCTV systems and the implementation of smart solutions and green technology as part of the Penang 2030 vision, aligning with the Sustainable Development Goals (SDGs). However, despite these efforts, road accidents persist due to non-optimised detection models, incomplete data from manual reporting, and technological constraints in real-time video analysis and predictive modelling. This study evaluates the effectiveness of the YOLOv7+G3HN framework for vehicle detection and near-miss analysis, with a focus on the influence of video quality on detection performance. The research aims to understand how high- and low-quality video inputs affect the accuracy and computational efficiency of detection algorithms. High[1]quality videos resulted in significantly faster computation times for vehicle detection than low-quality videos, highlighting the importance of video resolution in optimising detection processes. Despite the robustness of the algorithm, with no errors detected in both video qualities, higher miss detection rates in low-quality videos suggest that lower resolution may compromise detection accuracy and the effectiveness of monitoring systems. Near-miss analysis revealed that high-quality videos had a lower probability of near-miss occurrences than low-quality videos, highlighting the importance of video resolution for detection efficacy. These findings emphasise the critical role of high-resolution video inputs in enhancing detection accuracy and reliability, advocating for their implementation to optimise vehicle detection and improve road safety. Additionally, YOLOv7+G3HN outperforms YOLOv7 in both accuracy and speed. The study concludes that the YOLOv7+G3HN framework is effective for vehicle detection and near-miss analysis, provided that video quality is considered in system design and implementation.

Efficient Path Planning Algorithm Based on Laser SLAM and an Optimized Visibility Graph for Robots

Mobile robots’ efficient path planning has long been a challenging task due to the complexity and dynamism of environments. If an occupancy grid map is used in path planning, the number of grids is determined by grid resolution and the size of the actual environment. Excessively high resolution increases the number of traversed grid nodes and thus prolongs path planning time. To address this challenge, this paper proposes an efficient path planning algorithm based on laser SLAM and an optimized visibility graph for mobile robots, which achieves faster computation of the shortest path using the optimized visibility graph. Firstly, the laser SLAM algorithm is used to acquire the undistorted LiDAR point cloud data, which are converted into a visibility graph. Secondly, a bidirectional A* path search algorithm is combined with the Minimal Construct algorithm, enabling the robot to only compute heuristic paths to the target node during path planning in order to reduce search time. Thirdly, a filtering method based on edge length and the number of vertices of obstacles is proposed to reduce redundant vertices and edges in the visibility graph. Additionally, the bidirectional A* search method is implemented for pathfinding in the efficient path planning algorithm proposed in this paper to reduce unnecessary space searches. Finally, simulation and field tests are conducted to validate the algorithm and compare its performance with classic algorithms. The test results indicate that the method proposed in this paper exhibits superior performance in terms of path search time, navigation time, and distance compared to D* Lite, FAR, and FPS algorithms.

PyTWMR: transcriptome-wide Mendelian randomization in python.

Mendelian randomization (MR) is a widely used approach to estimate causal effect of variation in gene expression on complex traits. Among several MR-based algorithms, transcriptome-wide summary statistics-based Mendelian Randomization approach (TWMR) enables the uses of multiple SNPs as instruments and multiple gene expression traits as exposures to facilitate causal inference in observational studies. Here we present a Python-based implementation of TWMR and revTWMR. Our implementation offers GPU computational support for faster computations and robust computation mode resilient to highly correlated gene expressions and genetic variants. pyTWMR is available at github.com/soreshkov/pyTWMR.

Fast and scalable ensemble learning method for versatile polygenic risk prediction

Polygenic risk scores (PRS) enhance population risk stratification and advance personalized medicine, but existing methods face several limitations, encompassing issues related to computational burden, predictive accuracy, and adaptability to a wide range of genetic architectures. To address these issues, we propose Aggregated L0Learn using Summary-level data (ALL-Sum), a fast and scalable ensemble learning method for computing PRS using summary statistics from genome-wide association studies (GWAS). ALL-Sum leverages a L0L2 penalized regression and ensemble learning across tuning parameters to flexibly model traits with diverse genetic architectures. In extensive large-scale simulations across a wide range of polygenicity and GWAS sample sizes, ALL-Sum consistently outperformed popular alternative methods in terms of prediction accuracy, runtime, and memory usage by 10%, 20-fold, and threefold, respectively, and demonstrated robustness to diverse genetic architectures. We validated the performance of ALL-Sum in real data analysis of 11 complex traits using GWAS summary statistics from nine data sources, including the Global Lipids Genetics Consortium, Breast Cancer Association Consortium, and FinnGen Biobank, with validation in the UK Biobank. Our results show that on average, ALL-Sum obtained PRS with 25% higher accuracy on average, with 15 times faster computation and half the memory than the current state-of-the-art methods, and had robust performance across a wide range of traits and diseases. Furthermore, our method demonstrates stable prediction when using linkage disequilibrium computed from different data sources. ALL-Sum is available as a user-friendly R software package with publicly available reference data for streamlined analysis.

Derivation of Analytical Expressions for Fast Calculation of Resistance Spot Welding System Currents

Derivation of Analytical Expressions for Fast Calculation of Resistance Spot Welding System Currents

High-Definition Dynamic Voltage Restorer Systems Using Equivalent Time Sampling Techniques and Circular Structural Memory Filters

Due to advances in power electronics technology and the evolution of automation devices, the number of electrical devices that are sensitive to power quality is rapidly increasing, and for this reason, users are increasing their demand for high quality. To meet power quality demands, many power conversion devices are used, including dynamic voltage restorers (DVRs). DVRs are recognized as devices that can effectively manage problems such as voltage segments, swells, and harmonics. DVR control requires many samples for harmonic compensation, which has the disadvantage of being complicated to implement due to fast digital signal processing computation and the application of the cyclic discrete Fourier transform. In this paper, a high-precision DVR system configuration is proposed that compensates for harmonics using a periodically equivalent time-interval sampling technique and a novel circular-structured memory filter. The proposed circular-structured multi-pointer memory filter is an effective filter algorithm for high-precision input voltage measurement because it can remove noise and compensate for the delay of the phase angle of the filter in voltage measurement. A simulation and DVR prototype system were built, and the feasibility and effectiveness of the phase angle multi-filter voltage detection method and the compensation method were verified by experiments.

A python library for the fast and scalable computation of biologically meaningful individual specific networks

Individual Specific Networks (ISNs) are a tool used in computational biology to infer Individual Specific relationships between biological entities from omics data. ISNs provide insights into how the interactions among these entities affect their respective functions. To address the scarcity of solutions for efficiently computing ISNs on large biological datasets, we present ISN-tractor, a data-agnostic, highly optimized Python library to build and analyse ISNs. ISN-tractor demonstrates superior scalability and efficiency in generating Individual Specific Networks (ISNs) when compared to existing methods such as LionessR, both in terms of time and memory usage, allowing ISNs to be used on large datasets. We show how ISN-tractor can be applied to real-life datasets, including The Cancer Genome Atlas (TCGA) and HapMap, showcasing its versatility. ISN-tractor can be used to build ISNs from various -omics data types, including transcriptomics, proteomics, and genotype arrays, and can detect distinct patterns of gene interactions within and across cancer types. We also show how Filtration Curves provided valuable insights into ISN characteristics, revealing topological distinctions among individuals with different clinical outcomes. Additionally, ISN-tractor can effectively cluster populations based on genetic relationships, as demonstrated with Principal Component Analysis on HapMap data.

Evaluation of Traffic Sign Occlusion Rate Based on a 3D Point Cloud Space

The effectiveness of road signs is hindered by obstructions, such as vegetation, mutual obstruction of signs, or the road alignment itself. The traditional evaluation of road sign recognition effectiveness is conducted through in-vehicle field surveys. However, this method has several drawbacks, including discontinuous identification, unclear positioning, incomplete coverage, and being time-consuming. Consequently, it is unable to effectively assess the recognition status of road signs at any arbitrary point within the road space. Therefore, this study employed laser scanning to construct a point-surface model, which was based on a point cloud algorithm and SLAM (Simultaneous Localization and Mapping), integrated LiDAR and inertial navigation system data, and optimized the point model after processing steps such as denoising, resampling, and three-dimensional model construction. Furthermore, a method for assessing the highway sign occlusion rate based on the picking algorithm was proposed. The algorithm was applied to an actual road environment, and the occlusion by other items was simulated. The results demonstrated the effectiveness of the method. This new method provides support for the fast and accurate calculation of road sign occlusion rates, which is of great importance for ensuring the safe traveling of vehicles.

Computer-assisted regional-residual gravity anomaly separation with regularized first- and second-order derivatives

The regional-residual separation of gravity anomalies in crustal and mineral exploration was a graphical-based procedure before the advent of fast digital computers and the need for more efficient algorithms to process large data sets. However, since requiring the supervision of an experienced interpreter, the results, once obtained with graphical procedures, are often accepted as second to none in producing anomalies with geologic significance. Numerical methods based on spectral filtering and robust polynomial fitting have worked in many scenarios but seem not fully effective in replicating in algorithms the kind of results once obtained with interpreter-assisted graphical methods. We develop a computer-assisted regional residual separation (CARRS) procedure implemented by a set of short MATLAB scripts; in many aspects, CARRS simulates the operations of former graphical methods but requires few decisions and minor handwork from the interpreter. CARRS applies robust polynomial fitting to points with a low horizontal gradient and a vertical second-order derivative; thus, selecting fitting points as a real interpreter would do with graphical approaches in outlining the regional field. A regularized procedure is used to calculate the stable first- and second-order derivatives. Companion MATLAB codes allow for the replication of our results and further exploration with different threshold levels to identify flat domain regions. CARRS is illustrated with airborne gravity data CPRM-1123, freely available to download. A data window of the residual field is used to analyze the distribution of cassiterite deposits in greisen zones of the Paleoproterozoic Velho Guilherme granites in the Amazon Craton, as their distribution appears in the observed gravity anomaly and its corresponding residual field.

An automated lightweight approach for detecting dead fish in a recirculating aquaculture system

Accurate methods for detecting dead fish are of great significance for fisheries because of their potential to improve production and reduce water pollution. Small-scale fisheries typically apply manual observation, but this approach is labor-intensive and time-consuming, which limits its application on large-scale farms. Advances in computer vision tools have provided the foundation for revolutionary methods of automated identification of individual-animal behavior. However, computer vision systems often suffer from unstable accuracy, low efficiency, and limited detection capability. Thus, the aim of this study was to design a deep learning system, designated “Deadfish-YOLO”, to detect dead fish based solely on standard underwater camera images with no additional hardware. First, eight selected rearing tanks were monitored by a self-developed computer version system. From these tank images, a dataset containing 18,114 frames was built by manual labeling. Next, a lightweight backbone network was generated with YOLOv4 to ensure fast computations, and an attention mechanism was introduced into the model to suppress unimportant features. Finally, the ReLU-memristor-like activation function was adopted to improve neural-network performance. The accuracy and processing speed of Deadfish-YOLO were superior to those of other state-of-the-art single- and two-stage detection models; Deadfish-YOLO ran at 85 frames per second with a mean average precision of 0.946 and an average ratio of 0.924 between intersection and union. These results demonstrate that Deadfish-YOLO may be used to automatically monitor dead fish in real circulating aquaculture systems. In addition, the results of this study should facilitate the wider application of artificial-intelligence-based animal monitoring in aquaculture.

Fast computation of function composition derivatives for flatness-based control of diffusion problems

The chain rule is a standard tool in differential calculus to find derivatives of composite functions. Faà di Bruno’s formula is a generalization of the chain rule and states a method to find high-order derivatives. In this contribution, we propose an algorithm based on Faà di Bruno’s formula and Bell polynomials (Bell in Ann Math 29:38–46, 1927; Parks and Krantz in A primer of real analytic functions, 2012) to compute the structure of derivatives of function compositions. The application of our method is showcased using trajectory planning for the heat equation (Laroche et al. in Int J Robust Nonlinear Control 10(8):629–643, 2000).

High-Order Block Toeplitz Inner-Bordering method for solving the Gelfand–Levitan–Marchenko equation

We propose a high precision algorithm for solving the Gelfand–Levitan–Marchenko equation. The algorithm is based on the block version of the Toeplitz Inner-Bordering algorithm of Levinson’s type. To approximate integrals, we use the high-precision one-sided and two-sided Gregory quadrature formulas. Also we use the Woodbury formula to construct a computational algorithm. This makes it possible to use the almost Toeplitz structure of the matrices for the fast calculations. To the best of our knowledge, this is the first algorithm to solve this problem with an order of accuracy higher than the second.

Fast and Scaled Counting-Based Stochastic Computing Divider Design

Fast and Scaled Counting-Based Stochastic Computing Divider Design