Training Capabilities Research Articles

Effect of intramyocardial injection of allogeneic biomaterial on the exercise tolerance after destructive injury to the heart muscle (experimental study)

Introduction. The low regenerative potential of the heart prevents replacing dead cardiomyocytes with new cells. It has been shown that cell-free allogeneic biodegradable material (ABM) can increase the efficiency of reparative processes in myocardium and improve its structural and functional state. However, how the physical training capabilities change against this background has not been practically studied. Aim. To evaluate exercise tolerance in rats after the use of ABM. Material and methods. The work was performed on Wistar rats. Myocardial damage was caused by coronary occlusion (CO) and cryodestruction (CD) of the left ventricle apex of the heart. ABM was injected into the CO zone and along the perimeter of the myocardium affected area during CD. Animals with myocardial damage were included in control groups. Rats with myocardial damage and ABM injection constituted the experimental groups. Rats were taken into the experiment 45 days after ABM injection. The physical tolerance of rats was measured using the forced swimming test. Results. It has been shown that the swimming time of rats with СO by 62.5 % and with CD by 37.5 % (p<0.05) is less compared to similar indicators for these animals in the initial (intact) state, which indicates a significant decrease in their exercise tolerance in myocardial damage. Animals with CO or CD on the background of intramyocardial injection of ABM showed 33 % and 48 % longer swimming during the test, respectively. The increase in exercise tolerance of rats against the background of the ABM administration was accompanied by the improvement in the morphometric parameters of the heart. Conclusion. The effect of ABM intramyocardial injection on the endogenous mechanisms of cardiac muscle regeneration is effective and functionally significant.

ClickSAM: Fine-tuning Segment Anything Model using click prompts for ultrasound image segmentation.

The newly released Segment Anything Model (SAM) is a popular tool used in image processing due to its superior segmentation accuracy, variety of input prompts, training capabilities, and efficient model design. However, its current model is trained on a diverse dataset not tailored to medical images, particularly ultrasound images. Ultrasound images tend to have a lot of noise, making it difficult to segment out important structures. In this project, we developed ClickSAM, which fine-tunes the Segment Anything Model using click prompts for ultrasound images. ClickSAM has two stages of training: the first stage is trained on single-click prompts centered in the ground-truth contours, and the second stage focuses on improving the model performance through additional positive and negative click prompts. By comparing the first stage's predictions to the ground-truth masks, true positive, false positive, and false negative segments are calculated. Positive clicks are generated using the true positive and false negative segments, and negative clicks are generated using the false positive segments. The Centroidal Voronoi Tessellation algorithm is then employed to collect positive and negative click prompts in each segment that are used to enhance the model performance during the second stage of training. With click-train methods, ClickSAM exhibits superior performance compared to other existing models for ultrasound image segmentation.

The use of META (virtual simulations) in Canadian junior military leadership development

The use of META (virtual simulations) in Canadian junior military leadership development

Adaptive Gait Training of a Lower Limb Rehabilitation Robot Based on Human-Robot Interaction Force Measurement.

The existing fixed gait lower limb rehabilitation robots perform a predetermined walking trajectory for patients, ignoring their residual muscle strength. To enhance patient participation and safety in training, this paper aims to develop a lower limb rehabilitation robot with adaptive gait training capability relying on human-robot interaction force measurement. Firstly, a novel lower limb rehabilitation robot system with several active and passive driven joints is developed, and 2 face-to-face mounted cantilever beam force sensors are employed to measure the human-robot interaction forces. Secondly, a dynamic model of the rehabilitation training robot is constructed to estimate the driven forces of the human lower leg in a completely passive state. Thereafter, based on the theoretical moment from the dynamics and the actual joint interaction force collected by the sensors, an adaptive gait adjustment method is proposed to achieve the goal of adapting to the wearer's movement intention. Finally, interactive experiments are carried out to validate the effectiveness of the developed rehabilitation training robot system. The proposed rehabilitation training robot system with adaptive gaits offers great potential for future high-quality rehabilitation training, e.g., improving participation and safety.

An assessment of barriers to digital transformation in circular Construction: An application of stakeholder theory

The construction industry faces the rising challenge of circularity, and digital transformation emerges as a pivotal solution. However, limited stakeholder involvement impedes the adoption of digital transformation. The existing literature does not analyze stakeholders' perceptions of digitalization adoption barriers. Therefore, this study investigates the barriers hindering digital transformation in circular construction by examining the varied perceptions of key stakeholders. The study contributes to aligning stakeholders' interests and leads to targeted and effective interventions, enhancing the likelihood of successful adoption. Using the Stakeholder Theory, the study identified contractors, investors, clients, waste managers and government as the most influential stakeholders in circular construction. This study explores skill and knowledge barriers, market barriers, regulatory barriers, organizational and cultural barriers, and financial and resource barriers as key dimensions hindering digital circular practices. The study integrates the fuzzy Delphi method, fuzzy decision-making trial and evaluation laboratory, and analytical network processing to address uncertainties and interrelationships among barriers. The findings indicate that stakeholders hold differing perceptions regarding the most critical barriers; however, there is a consensus that critical aspects are skill and knowledge, organizational and cultural, and regulatory barriers. Practical barriers include lack of government support, resistance to change, lack of feasibility information, lack of training capabilities, and poor digital literacy. The study provides theoretical and practical implications, informing future research and interventions to develop targeted strategies and initiatives.

Unsupervised Action Segmentation via Fast Learning of Semantically Consistent Actoms

Action segmentation serves as a pivotal component in comprehending videos, encompassing the learning of a sequence of semantically consistent action units known as actoms. Conventional methodologies tend to require a significant consumption of time for both training and learning phases. This paper introduces an innovative unsupervised framework for action segmentation in video, characterized by its fast learning capability and absence of mandatory training. The core idea involves splitting the video into distinct actoms, which are then merging together based on shared actions. The key challenge here is to prevent the inadvertent creation of singular actoms that attempt to represent multiple actions during the splitting phase. Additionally, it is crucial to avoid situations where actoms associated with the same action are incorrectly grouped into multiple clusters during the merging phase. In this paper, we present a method for calculating the similarity between adjacent frames under a subspace assumption. Then, we employ a local minimum searching procedure, which effectively splits the video into coherent actoms aligned with their semantic meaning and provides us an action segmentation proposal. Subsequently, we calculate a spatio-temporal similarity between actoms, followed by developing a merging process to merge actoms representing identical actions within the action segmentation proposals. Our approach is evaluated on four benchmark datasets, and the results demonstrate that our method achieves state-of-the-art performance. Besides, our method also achieves the optimal balance between accuracy and learning time when compared to existing unsupervised techniques. Code is available at https://github.com/y66y/SaM.

Mimicking Bidirectional Inhibitory Synapse Using a Porous-Confined Ionic Memristor with Electrolyte/Tris(4-aminophenyl)amine Neurotransmitter.

Ionic memristors can emulate brain-like functions of biological synapses for neuromorphic technologies. Apart from the widely studied excitatory-excitatory and excitatory-inhibitory synapses, reports on memristors with the inhibitory-inhibitory synaptic behaviors remain a challenge. Here, the first biaxially inhibited artificial synapse is demonstrated, consisting of a solid electrolyte and conjugated microporous polymers bilayer as neurotransmitter, with the former serving as an ion reservoir and the latter acting as a confined transport. Due to the migration, trapping, and de-trapping of ions within the nanoslits, the device poses inhibitory synaptic plasticity under both positive and negative stimuli. Remarkably, the artificial synapse is able to maintain a low level of stable nonvolatile memory over a long period of time (≈60min) after multiple stimuli, with feature-inferencing/-training capabilities of neural node in neuromorphic computing. This work paves a reliable strategy for constructing nanochannel ionic memristive materials toward fully inhibitory synaptic devices.

The Effect of Innovation and Training Capabilities on the Performance of Village Equipment in Palimanan District

Human resources are a crucial factor in the development, progress, and change process in both organizations and government agencies. Mainly to improve human welfare productively. Palimanan District is a sub-district in Cirebon Regency. Palimanan District has 12 villages led by the village head and on the road or assisted by village officials and staff in carrying out duties and authorities. The ability to innovate is essential for the village to increase its potential in the village. The study aimed to determine the effect of innovation and training capabilities on the performance of village apparatus in the Palimanan sub-district by applying observation models, data collection, and questionnaire models in villages and Palimanan sub-districts. The research design used in this study was a descriptive analysis with a quantitative approach, aiming to determine the relationship between two or more variables. The instrument used for this study was a questionnaire sheet. This is shown from the results of the T-Test, which shows the calculated T value for the Inovas ability variable 2.431 and T table 1989, for the training variable of 5.030 and T table 1989. So it can be concluded that the calculated T value is greater than the T table where the innovation ability variable affects performance and the training variable affects performance; the F value is calculated at 35.365 > Ftable 3.11, which means there is a significant favorable influence between innovation ability and training together (simultaneously) on the performance of Palimanan sub-district village apparatus.

AFP Intelligence Community’s Strategic Risk Management Process in Response to External Aggression

In the ever-changing global security landscape, characterized by volatility, uncertainty, complexity, and ambiguity (VUCA), this study serves as a capstone project to provide valuable insights to the AFP Intelligence Community (AFPIC). Its primary focus is to effectively tackle the exceptional non-military security threats a foreign adversary poses. China’s rapid and significant ascent as an Asian hegemon and a global superpower has driven its pursuit of resources and influence worldwide. These hegemonic actions have not only elicited resentment from smaller and weaker nations adversely affected by them but also drawn the ire of the United States and its allies. The Philippines, strategically located in the South China Sea and the Pacific region, has become a prime target of China’s unconventional warfare strategy, where they seek to subdue their adversaries without resorting to kinetic force. China’s United Front Works (UFW) strategy is a crucial element of its hybrid warfare approach, providing a platform for intelligence, influence, and interference operations designed to systematically undermine its enemies. This study aims to describe the current state of the AFPIC’s risk management capabilities in terms of doctrines, organization, and training concerning the security threats posed by China’s UFW. It also seeks to identify the challenges in these areas and propose appropriate solutions to address them. Overall, the objective of this study is to contribute to the long-term endeavor of enhancing the AFPIC’s doctrines, organization, and training to effectively address the threats posed by China’s UFW and other hybrid security threats. The qualitative data gathered through document analysis, key informant interviews, and focused group discussions involving multiple participants from various units and offices within the AFPIC and the AFP Triad, as well as education and doctrine development circles, underwent an iterative and hybrid process of thematic and content analysis. Tools such as SWOT, TOWS, and the causal-impact analytical tool were utilized during this analytical process. The results of this analysis provided valuable insights into the current state of AFPIC’s risk management capability in terms of doctrine, organization, and training, highlighting both the advantages and disadvantages within these areas. Further exploration revealed the primary challenges present in these three domains, which encompassed internal and external dynamics that hindered the optimal implementation of related initiatives. Subsequently, numerous recommendations were formulated to enhance the AFPIC’s doctrine, organization, and training capabilities in response to the exceptional security threats posed by China’s united front works. Central to the research findings were three crucial factors that influenced the success or failure of the doctrinal, organizational, and training endeavors: the support of advocates and champions, the role of knowledge management, and the significance of engaging foreign stakeholders. Keywords: Armed Forces of the Philippines, Intelligence, Risk Management, United Front Works, National Security, hybrid security threats, unconventional warfare, political warfare

ChatGPT Makes a Difference in Daily Life: As a Life Guidance

Although Chat Generative Pre-trained Transformer (ChatGPT) has made significant strides in various domains such as education, healthcare, and music, the current situation highlights a lack of comprehensive research and analysis regarding its influence on these vital aspects of life. This research commences by outlining ChatGPT's training capabilities and its developmental journey. Subsequently, this review delve into its practical application in real-world scenarios, where ChatGPT is assigned the task of generating a 90-second background score while adhering to specific constraints. A notable assessment by Marco Cascella and colleagues is then presented, where ChatGPT was engaged to generate a medical report for ICU patients, using a set of non-sequentially provided medical parameters such as treatments, laboratory results, and respiratory metrics. This section also critically examines the limitations of ChatGPT in handling medical queries. Further, this review discussed the generative AI architecture of Baidu Anu, D, highlighting its remarkable efficacy, using ChatGPT as an illustrative example, particularly in pedagogical settings. The paper concludes by acknowledging the substantial contributions of ChatGPT in various arenas, while also addressing its inherent limitations. The study findings suggest that ChatGPT can notably benefit areas such as education, healthcare, or music, thanks to the robust language AI model created by OpenAI.

Microvibration Control of Hybrid Platform for High-Tech Equipment in Buildings Using Brain Emotional Learning-Based Intelligent Controller

In the semiconductor industry, the vulnerability of high-tech facilities installed on platforms to ground excitations induced by nearby traffic is significantly pronounced, primarily due to their small-scale dimensions. Consequently, it is imperative to design a smart control technique by effectively utilizing model-free controllers. Recently, adaptive intelligent control algorithms have emerged as a viable alternative to conventional model-based control algorithms. To address this issue, this study meticulously designed a hybrid platform using the adaptive intelligent controller known as the brain emotional learning-based intelligent controller, along with a Sugano fuzzy inference system to effectively optimize the controller’s learning parameters. These learning and intelligent-based algorithms offer notable advantages, including the ability to handle nonlinearity, uncertainty, and training capabilities within the control systems. To assess the effectiveness of the proposed controller in mitigating vibrations induced by traffic on high-tech facilities, a three degree-of-freedom structure is employed along with the hybrid platform. Finally, the performance of the hybrid platform in terms of microvibration control levels is meticulously validated using the Bolt Beranek Newman vibration criteria. Simulation results unequivocally demonstrate that the proposed controller outperforms both an uncontrolled system and a traditional linear quadratic regulator controller in terms of reducing the traffic-induced response of the hybrid platform and second floor, respectively. Through the integration of learning and intelligent-based controllers, the velocity levels of both the hybrid platform and the second floor are reduced to approximately 49.02% and remain well within the acceptable standard criteria curves.

Fault detection and classification using deep learning method and neuro‐fuzzy algorithm in a smart distribution grid

AbstractThis article proposes a deep learning (DL) model made of Long Short Term Memory (LSTM) and Adaptive Neuro Fuzzy Inference System (ANFIS) to detect fault in smart distribution grid assisted by communication systems using smart meter data. In smart grid, data analysis for fault identification and detection is crucial for grid monitoring. Nowadays, there are several DL techniques developed for smart grid data analysis applications. To solve this problem, a novel data analysis model based on deep learning and Neuro‐fuzzy algorithm is developed for fault location in a smart power grid. First, the LSTM is applied for training the data samples extracted from the smart meters. Then, an ANFIS algorithm is implemented for fault detection and identification from the trained data. Finally, faults are located with the higher accuracy. With this intelligent method proposed, single‐phase, two‐phase and three‐phase faults can be identified using a restricted amount of data. The novelty of the proposed method compared with other methods is the capability of fast training and testing even with large amount of data. To verify the effectiveness of our methodology, an intelligent model of the IEEE 13‐node network is used. The effectiveness and robustness of the proposed model are evaluated using several parameters such as accuracy, precision‐recall, F1‐score, Receiver Operating Characteristic (ROC) curve and complexity time. The obtained results indicate that the proposed deep learning model outperforms existing deep learning methods in the literature for fault detection and classification with 99.99% accuracy.

Evolving reliability assessment of systems using active learning-based surrogate modelling

Risk tracking involves monitoring and assessing evolving event risks in complex infrastructure systems. As these systems change over time, so do the associated risk profiles, requiring reliability updating principles. Traditional methods rely on repetitive simulations of complex models, while recent surrogate-based techniques offer efficiency and accuracy without relying on observed data. However, these approaches inadequately address time-dependent reliability updating and surrogate modeling integration. To overcome these limitations, this study presents the Time-Variant Reliability Updating method (TVRU), which seamlessly integrates Kriging-based time-dependent reliability with parallel learning. TVRU estimates prior failure probability by leveraging the adaptive training capabilities of a time-variant surrogate Kriging model. This model incorporates random variables and time as inputs, using the performance function's responses as outputs. By classifying design samples, the prior failure probability is accurately computed, and a meticulous Kriging model estimates the posterior probability of failure. The TVRU method is demonstrated using four numeric models of varying complexities, showcasing exceptional efficiency and accuracy. This novel approach enhances risk assessment in complex systems, facilitating informed decision-making.

Open-loop analog programmable electrochemical memory array

Emerging memories have been developed as new physical infrastructures for hosting neural networks owing to their low-power analog computing characteristics. However, accurately and efficiently programming devices in an analog-valued array is still largely limited by the intrinsic physical non-idealities of the devices, thus hampering their applications in in-situ training of neural networks. Here, we demonstrate a passive electrochemical memory (ECRAM) array with many important characteristics necessary for accurate analog programming. Different image patterns can be open-loop and serially programmed into our ECRAM array, achieving high programming accuracies without any feedback adjustments. The excellent open-loop analog programmability has led us to in-situ train a bilayer neural network and reached software-like classification accuracy of 99.4% to detect poisonous mushrooms. The training capability is further studied in simulation for large-scale neural networks such as VGG-8. Our results present a new solution for implementing learning functions in an artificial intelligence hardware using emerging memories.

Understanding Cloud Systems’ Structure and Organization Using a Machine’s Self-Learning Approach

Abstract In this study, we introduce a self-supervised deep neural network approach to classify satellite images into independent classes of cloud systems. The driving question of the work is to understand whether our algorithm can capture cloud variability and identify distinct cloud regimes. Ultimately, we want to achieve generalization such that the algorithm can be applied to unseen data and thus help automatically extract relevant information important to atmospheric science and renewable energy applications from the ever-increasing satellite data stream. We use cloud optical depth (COD) retrieved from postprocessed high-resolution Meteosat Second Generation (MSG) satellite data as input for the network. The network’s architecture is based on the DeepCluster, version 2, and consists of a convolutional neural network and a multilayer perceptron, followed by a k-means algorithm. We explore the network’s training capabilities by analyzing the centroids and feature vectors found from progressive minimization of the cross-entropy loss function. By making use of additional MSG retrieval products based on multichannel information, we derive the optimum number of classes to determine independent cloud regimes. We test the network capabilities on COD data from 2013 and find that the trained neural network gives insights into the cloud systems’ persistence and transition probability. The generalization on the 2015 data shows good skills of our algorithm with unseen data, but results depend on the spatial scale of cloud systems. Significance Statement This study uses a self-supervised deep neural network to identify distinct cloud systems from cloud optical depth satellite images over central Europe. Satellite-retrieved products support the physical interpretation of the identified cloud classes and help optimize the number of identified classes. The trained neural network gives insights into cloud systems’ persistence and transition probability. The generalization capacity of the deep neural network with unseen data is promising but depends on the spatial scale of cloud systems.

Effects of Mg Doping to a LiCoO2 Channel on the Synaptic Plasticity of Li Ion-Gated Transistors.

Artificial synapses with ideal functionalities are essential in hardware neural networks to allow for energy-efficient analog computing. However, the realization of linear and symmetric weight updates in real synaptic devices has proven challenging and ultimately limits the online training capabilities of neural network systems. Herein, we investigate the effect of Mg doping on a LiCoO2 (LCO) channel in a Li ion-gated synaptic transistor, so as to improve long-term and short-term plasticity. Two transistor structures, based on a lithium phosphorus oxynitride electrolyte, were examined by using undoped LCO and Mg-doped LCO as the channel material between the source and drain electrodes. It was found that Mg doping increased the initial channel conductance by 3 orders of magnitude, which is probably due to the substitution of Co3+ by Mg2+ and the compensation of hole creation. It was further found that the doped channel transistor showed good retention characteristics and better linearity of long-term potentiation and depression when voltage pulses were applied to the gate electrode. The improved retention and linearity are attributed to an extended range of the insulator-to-conductor transition by Mg doping and Li-ion extraction/insertion cooperated in the LCO channel. Using the obtained synaptic weight update, artificial neural network simulations demonstrated that the doped channel transistor shows an image recognition accuracy of ∼80% for handwritten digits, which is higher than ∼65% exhibited by the undoped channel transistor. Mg doping also improved short-term plasticity such as paired-pulse facilitation/depression and Hebbian spike timing-dependent plasticity. These results indicate that elemental doping to the channel of Li ion-gated synaptic transistors could be a useful procedure for realizing robust neuromorphic systems based on analog computing.

Exploiting UAV for Air–Ground Integrated Federated Learning: A Joint UAV Location and Resource Optimization Approach

Recently, many exciting usage scenarios and groundbreaking technologies for sixth generation (6G) networks have drawn more and more attention. The revolution of 6G mainly lies in ubiquitous intelligence, which promotes the development of edge intelligence (EI) by running artificial intelligence (AI) algorithms at the network edge. By embedding training capabilities across the network nodes, federated learning (FL) can achieve high security and alleviate network traffic congestion, which provides a promising way to realize the ubiquitous EI. While traditional FL usually relies static terrestrial base stations (BSs) for the global model aggregation, unmanned aerial vehicles (UAVs) could effectively supplement the terrestrial BSs because of their high maneuverability, thereby building the air-ground integrated FL (AGIFL). Nevertheless, how to effectively deploy the UAV and allocate resources to boost the learning performance and achieve high energy efficiency in the AGIFL remains largely unexplored. In this paper, we study how to jointly optimize the UAV location and resource allocation to minimize the incurred cost in terms of two objectives: i) the minimization of terrestrial users’ energy consumption; ii) the minimization of tradeoff between energy consumption and training latency. The formulated non-convex problems are efficiently solved by alternating optimization techniques based on successive convex approximation (SCA) approaches after appropriate problem decomposition. Extensive simulation results show that our proposed algorithms can reduce more cost than three benchmarks while guaranteeing the learning accuracy. Furthermore, we construct a real-world AGIFL system, implement the proposed algorithms in the system, and carry out field experiments to verify the superiority of our algorithms.

A life-threatening bleeding prediction model for immune thrombocytopenia based on personalized machine learning: a nationwide prospective cohort study

Rare but critical bleeding events in primary immune thrombocytopenia (ITP) present life-threatening complications in patients with ITP, which severely affect their prognosis, quality of life, and treatment decisions. Although several studies have investigated the risk factors related to critical bleeding in ITP, large sample size data, consistent definitions, large-scale multicenter findings, and prediction models for critical bleeding events in patients with ITP are unavailable. For the first time, in this study, we applied the newly proposed critical ITP bleeding criteria by the International Society on Thrombosis and Hemostasis for large sample size data and developed the first machine learning (ML)-based online application for predict critical ITP bleeding. In this research, we developed and externally tested an ML-based model for determining the risk of critical bleeding events in patients with ITP using large multicenter data across China. Retrospective data from 8 medical centers across the country were obtained for model development and prospectively tested in 39 medical centers across the country over a year. This system exhibited good predictive capabilities for training, validation, and test datasets. This convenient web-based tool based on a novel algorithm can rapidly identify the bleeding risk profile of patients with ITP and facilitate clinical decision-making and reduce the occurrence of adversities.

Electrically programmable phase-change photonic memory for optical neural networks with nanoseconds in situ training capability

Electrically programmable phase-change photonic memory for optical neural networks with nanoseconds in situ training capability

Towards Adversarial Robustness for Multi-Mode Data through Metric Learning.

Adversarial attacks have become one of the most serious security issues in widely used deep neural networks. Even though real-world datasets usually have large intra-variations or multiple modes, most adversarial defense methods, such as adversarial training, which is currently one of the most effective defense methods, mainly focus on the single-mode setting and thus fail to capture the full data representation to defend against adversarial attacks. To confront this challenge, we propose a novel multi-prototype metric learning regularization for adversarial training which can effectively enhance the defense capability of adversarial training by preventing the latent representation of the adversarial example changing a lot from its clean one. With extensive experiments on CIFAR10, CIFAR100, MNIST, and Tiny ImageNet, the evaluation results show the proposed method improves the performance of different state-of-the-art adversarial training methods without additional computational cost. Furthermore, besides Tiny ImageNet, in the multi-prototype CIFAR10 and CIFAR100 where we reorganize the whole datasets of CIFAR10 and CIFAR100 into two and ten classes, respectively, the proposed method outperforms the state-of-the-art approach by 2.22% and 1.65%, respectively. Furthermore, the proposed multi-prototype method also outperforms its single-prototype version and other commonly used deep metric learning approaches as regularization for adversarial training and thus further demonstrates its effectiveness.