Computational Infrastructure Research Articles

Impact of Tracheostomy Timing Within the National Veterans Affairs Population.

There is a lack of a definitive study in the literature comparing early versus late tracheostomy and exploring the impact of tracheostomy timing on patient outcomes. This study may help guide treatment paradigms and contribute to a consensus for optimal tracheostomy timing. A retrospective review was performed comparing early versus late timing of tracheostomy placement and their respective outcomes. The authors used data provided by VA Informatics and Computing Infrastructure (VINCI) to find patients who received a tracheostomy at any VA Medical Center in the United States. There were a total of 25,334 tracheostomies in the database which satisfied our criteria. These occurred between the years 1999 and 2022. Propensity score matching assessed 17,074 tracheostomies, 8537 in either group. The median age of patients in the matched groups was 66 years, and approximately 97.4% of patients were male. Early tracheostomy timing was defined as the placement of the tracheostomy within 10 days of intubation. Outcomes included post-tracheostomy intensive care unit (ICU) days, post-tracheostomy hospital days, successful ventilator weaning, and all-cause mortality. Early tracheostomy was associated with significantly fewer ICU days and hospital days, and the early group experienced higher rates of successful ventilator weaning. Survival analysis of data within 5 years of tracheostomy showed that early tracheostomy was associated with significantly lower hazard for all-cause mortality. Our results add to the body of evidence that an earlier transition to mechanical ventilation by tracheostomy confers benefits in patient morbidity and mortality as well as resource utilization. 3 Laryngoscope, 134:3555-3561, 2024.

Dual-drive coordinated operation strategy for internet data centers by exploiting spatiotemporal flexibility

With the rapid expansion of computational infrastructure, the electricity consumption of geographically dispersed Internet Data Centers (IDCs) has imposed a substantial influence on the power grid. Geographically distributed IDCs, which present a versatile electricity demand with significant operational flexibility, hold potential for mitigating cost issues and balancing regional supply and load demand. To exploit this flexibility, this paper has developed a dual-drive coordinated operational strategy, which integrates locational electricity prices and unit carbon emission costs based on the dynamic carbon emission intensity (DCEI) of various regions as interaction signals to improve coordination effectiveness. The upper-level determines the consumption behaviour of IDCs, the lower-level provides feedback on dual-drive signals factoring in IDCs' involvement in grid operation, whilst an iterative algorithm is suggested to solve the bi-level model. Case studies based on real-world datasets demonstrate that the proposed operation strategy can achieve arbitrage through multi-regional carbon-electricity markets, significantly reduce operating costs, and support the rapid development of IDCs. Additionally, this strategy facilitates optimal matching of power load in IDCs with renewable energy sources, thereby mitigating imbalances in regional renewable energy supply and demand.

CLOUD BASED DUPLICATION REMOVAL SYSTEM

Deduplication involves eliminating duplicate or redundant data to reduce stored data volume, commonly used in data backup, network optimization, and storage management. However, traditional deduplication methods have limitations with encrypted data and security. The primary objective of this project is to develop new distributed deduplication systems that offer increased reliability. In these systems, data chunks are distributed across the Hadoop Distributed File System (HDFS), and a robust key management system is utilized to ensure secure deduplication with slave nodes. Instead of having multiple copies of the same content, deduplication removes redundant data by retaining only one physical copy and referring other instances to that copy. The granularity of deduplication can vary, ranging from an entire file to a data block. The MD5 and 3DES algorithms are used to enhance the deduplication process. The proposed approach in this project is the Proof of Ownership (POF) of the file. With this method, deduplication can effectively address the issues of reliability and label consistency in HDFS storage systems. The proposed system has successfully reduced the cost and time associated with uploading and downloading data, while also optimizing storage space. Key Words: Cloud computing, data storage, file checksum algorithms, computational infrastructure, duplication.

AI-Assisted Enhanced Composite Metric-Based Intrusion Detection System for Secured Cyber Internet Security for Next-Generation Wireless Networks

The prevalence of distributed denial of service (DDoS) flooding assaults is one of the most serious risks to cloud computing security. These types of assaults have as their primary objective the exhaustion of the system’s available resources, that is, the target of the attack, in order to make the system in question unavailable to authorized users. Internet thieves often conduct flooding assaults of the kind known as DDoS, focusing primarily on the application and network levels. When the computer infrastructure is multi-mesh-geo distributed, includes multi-parallel services, and a high number of domains, it may be difficult to detect assaults. This is particularly true when a substantial number of domains are present. When there are a big number of independent administrative users using the services, the situation gets more complicated. The purpose of this body of research is to identify signs that may be utilized to detect DDoS flooding assaults; this is its main objective. As a result, throughout the course of our study, we established a composite metric that considers application, system, network, and infrastructure elements as possible indicators of the incidence of DDoS assaults. According to our research, DDoS assaults may be triggered by a combination of variables. Investigations of simulated traffic are being conducted in the cloud. High traffic may be the result of flooding assaults. The composite metric-based intrusion detection system will be the name of a one-of-a-kind intrusion detection system (IDS) that has been agreed upon ICMIDS. This system will use [Formula: see text]-Means clustering and the Genetic Algorithm (GA) to detect whether an effort has been made to flood the cloud environment. CMIDS employs a multi-threshold algorithmic strategy in order to identify malicious traffic occurring on a cloud-based network. Cisco has created this technology. This strategy necessitates a comprehensive investigation of all factors, which is crucial for assuring the continuation of cloud-based computing-based activities. This monitoring system involves the development, administration, and storage of a profile database, denoted as Profile DB. This database is used for recording and using the composite metric for each virtual machine. The results of a series of tests are compared to the ISCX benchmark dataset and statistical settings. The results indicate that ICMIDS has a reasonably high detection rate and the lowest false alarm rate in the majority of situations examined during the series of tests done to validate and verify its efficacy. This was shown by the fact that ICMIDS had the lowest false alarm rate among all examined conditions.

Cloud media video encoding: review and challenges

AbstractIn recent years, Internet traffic patterns have been changing. Most of the traffic demand by end users is multimedia, in particular, video streaming accounts for over 53%. This demand has led to improved network infrastructures and computing architectures to meet the challenges of delivering these multimedia services while maintaining an adequate quality of experience. Focusing on the preparation and adequacy of multimedia content for broadcasting, Cloud and Edge Computing infrastructures have been and will be crucial to offer high and ultra-high definition multimedia content in live, real-time, or video-on-demand scenarios. For these reasons, this review paper presents a detailed study of research papers related to encoding and transcoding techniques in cloud computing environments. It begins by discussing the evolution of streaming and the importance of the encoding process, with a focus on the latest streaming methods and codecs. Then, it examines the role of cloud systems in multimedia environments and provides details on the cloud infrastructure for media scenarios. After doing a systematic literature review, we have been able to find 49 valid papers that meet the requirements specified in the research questions. Each paper has been analyzed and classified according to several criteria, besides to inspect their relevance. To conclude this review, we have identified and elaborated on several challenges and open research issues associated with the development of video codecs optimized for diverse factors within both cloud and edge architectures. Additionally, we have discussed emerging challenges in designing new cloud/edge architectures aimed at more efficient delivery of media traffic. This involves investigating ways to improve the overall performance, reliability, and resource utilization of architectures that support the transmission of multimedia content over both cloud and edge computing environments ensuring a good quality of experience for the final user.

User perception as a factor for improving Trustworthiness in e-commerce systems

The User Interface (UI) is the first artefact that the user interacts with while developing a sense of trust that motivates him to use software applications more effectively. A badly designed UI can deceive users and bring the system into disrepute. Trustworthiness in UI is mandatory, as a poorly implemented UI can lead to the user misusing the system and jeopardizing the expected result. Trust in computational systems involves not only technical aspects, such as computational infrastructure, storage space, and service composition, but also aspects of Human-Computer Interaction (HCI). While technological aspects have received considerable attention, there are few research on human-computer interaction in terms of trust. This paper describes a way for assessing a system’s trustworthiness based on the user’s perception. The approach relies on a quality model to aggregate interface quality criteria in order to get a trustworthiness score. Three sets of experiments involving more than 300 individuals were carried out to validate the suggested methodology. A comparison was made between the trustworthiness score obtained through the methodology and the answers to open questions obtained through the users’ questionnaires. The results were consistent, and statistical analysis corroborated the positive assessment. Based on these results, examples of improvements were developed to highlight the usefulness of the approach for developing more trustworthiness interfaces.

Security in Machine Learning (ML) Workflows

Purpose: This paper addresses the comprehensive security challenges inherent in the lifecycle of machine learning (ML) systems, including data collection, processing, model training, evaluation, and deployment. The imperative for robust security mechanisms within ML workflows has become increasingly paramount in the rapidly advancing field of ML, as these challenges encompass data privacy breaches, unauthorized access, model theft, adversarial attacks, and vulnerabilities within the computational infrastructure.
 Methodology: To counteract these threats, we propose a holistic suite of strategies designed to enhance the security of ML workflows. These strategies include advanced data protection techniques like anonymization and encryption, model security enhancements through adversarial training and hardening, and the fortification of infrastructure security via secure computing environments and continuous monitoring.
 Findings: The multifaceted nature of security challenges in ML workflows poses significant risks to the confidentiality, integrity, and availability of ML systems, potentially leading to severe consequences such as financial loss, erosion of trust, and misuse of sensitive information.
 Unique Contribution to Theory, Policy and Practice: Additionally, this paper advocates for the integration of legal and ethical considerations into a proactive and layered security approach, aiming to mitigate the risks associated with ML workflows effectively. By implementing these comprehensive security measures, stakeholders can significantly reinforce the trustworthiness and efficacy of ML applications across sensitive and critical sectors, ensuring their resilience against an evolving landscape of threats.

The effect of adaptive mesh refinement and grid stretching on the magnetized coronal mass ejection model in Icarus

Context. Coronal mass ejections (CMEs) are the main driver of solar wind disturbances near the Earth. When directed toward Earth, the internal magnetic field of the CME can interact with the Earth’s magnetic field and cause geomagnetic storms. In order to better predict and avoid damage coming from such events, the optimized heliospheric model Icarus has been implemented. Advanced numerical techniques, such as gradual radial grid stretching and solution adaptive mesh refinement (AMR) are implemented in the model to achieve better performance and more reliable results. Aims. The impact of a CME at Earth is greatly affected by its internal magnetic field structure. The aim of this work is to enable the modeling of the evolution of the magnetic field configuration of the CME throughout its propagation in Icarus. Thus, we used Icarus to implement a magnetized CME model that is more realistic than the already available simple hydrodynamics cone CME model, allowing us to study the evolution of the magnetized CME during its interactions with the solar wind. The focus of the study is on the global magnetic structure of the CME and its evolution and interaction with the solar wind. Methods. The magnetized CME model implemented in Icarus is the linear force-free spheromak (LFFS) solution that has been imported from EUHFORIA. Simulations with the spheromak model were performed for different effective resolutions of the computational domain. We applied advanced techniques such as grid stretching and AMR. Different AMR levels were applied in order to obtain high resolution locally, where needed. The original uniform medium- and high-resolution simulation results are also shown as a reference. The results of all the simulations are compared in detail and the wall-clock times of the simulations are provided. Results. We analyzed the results from the performed simulations. The co-latitudinal magnetic field component is plotted at 1 AU for both Icarus and EUHFORIA simulations. The time series at Earth (L1) of the radial velocity, density, and different magnetic field components are plotted and compared. The arrival time is better approximated by the EUHFORIA simulation, with the CME shock arriving 1.6 and 1.09 h later than in the AMR level 4 and 5 simulations, respectively. The profile features and variable strengths are best modeled by Icarus simulations with AMR level 4 and 5. The uniform, medium-resolution simulation with Icarus took 6.5 h wall-clock time, whereas with EUHFORIA, the most similar setup takes 18.5 h, when performed on 1 node with 2 Xeon Gold 6240 CPUs at 2.6 GHz (Cascadelake), 18 cores each, on the Genius cluster at KU Leuven. The Icarus simulation with AMR level 4 took only 2.5 h on the same computer infrastructure, while showing better resolved shocks and magnetic field features, when compared to the observational data and the referene uniform simulation results. Conclusions. The results from different Icarus simulations in Icarus are presented using results from the EUHFORIA heliospheric modeling tool as a reference. The arrival time is closer to the observed time in the EUHFORIA simulation, but the profiles of the different variables show more features and details in the Icarus simulations. The simulations with AMR levels 4 and 5 offered the most detailed results. Considering the small difference in the modeled results and the large difference in terms of computational resources, the AMR level 4 simulation is considered to have displayed the most optimal performance. The gradients in the AMR level 4 results are sharper than those in the uniform simulations with both EUHFORIA and Icarus, while the AMR level 4 effective resolution is the most comparable to the standard resolution runs. The AMR level 3 simulation is 15 and 41 times faster than the Icarus and EUHFORIA uniform simulations, respectively; while the AMR level 4 simulation is about three and seven times faster than the uniform simulations, respectively.

A ventriculomegaly feature computational pipeline to improve the screening of normal pressure hydrocephalus on CT.

The objective of this study was to develop a computational pipeline that extracts objective features of ventriculomegaly from non-contrast CT (NCCT) for the accurate classification of idiopathic normal pressure hydrocephalus (NPH) from headache controls (HCs), Alzheimer's dementia (AD), and posttraumatic encephalomalacia (PTE). Patients with possible NPH (n = 79) and a subset with definite NPH (DefNPH; n = 29) were retrospectively identified in the Veterans Affairs Informatics and Computing Infrastructure system, along with the AD (n = 62), PTE (n = 53), and HC (n = 59) cohorts. Image-processing pipelines were developed to extract a novel feature capturing the maximum eccentricity of the lateral ventricles (MaxEccLV), a proxy splenial angle (p-SA), the Evans indices (EI-x, -y, and -z), callosal angle, normalized maximum third-ventricle width, and CSF to brain volume ratio from their NCCT scans. The authors used t-tests to examine group differences in the features and multivariate logistic regression models for classification. Additionally, the NPH versus HC classifier was validated on external data. When NPH and DefNPH were compared with HC, AD, and PTE, significant differences were found in all features except the p-SA, which only significantly differed between NPH and PTE. The test-set area under the receiver operating characteristic curve (AUC), sensitivity, and specificity were 0.98, 100%, and 98.3% for NPH versus HC classification; 0.94, 87.3%, and 85.5% for NPH versus AD; 0.96, 92.4%, and 90.6% for NPH versus PTE; and 0.96, 94%, and 88% for NPH versus the other groups using logistic regression under five-fold cross-validation. Consistently high performance was noted for DefNPH. The NPH versus HC classifier provided an AUC of 0.84, sensitivity of 76.9%, and specificity of 90% when assessed on external data. Including the novel MaxEccLV, this framework computes useful features of ventriculomegaly, which had not previously been algorithmically assessed on NCCT. This framework successfully classified possible and definite NPH from HC, AD, and PTE. Following validation on larger representative cohorts, this objective and accessible tool may aid in screening for NPH and differentiating it from symptomatic mimics such as AD and PTE.

Membrane lipids drive formation of KRAS4b-RAF1 RBDCRD nanoclusters on the membrane

The oncogene RAS, extensively studied for decades, presents persistent gaps in understanding, hindering the development of effective therapeutic strategies due to a lack of precise details on how RAS initiates MAPK signaling with RAF effector proteins at the plasma membrane. Recent advances in X-ray crystallography, cryo-EM, and super-resolution fluorescence microscopy offer structural and spatial insights, yet the molecular mechanisms involving protein-protein and protein-lipid interactions in RAS-mediated signaling require further characterization. This study utilizes single-molecule experimental techniques, nuclear magnetic resonance spectroscopy, and the computational Machine-Learned Modeling Infrastructure (MuMMI) to examine KRAS4b and RAF1 on a biologically relevant lipid bilayer. MuMMI captures long-timescale events while preserving detailed atomic descriptions, providing testable models for experimental validation. Both in vitro and computational studies reveal that RBDCRD binding alters KRAS lateral diffusion on the lipid bilayer, increasing cluster size and decreasing diffusion. RAS and membrane binding cause hydrophobic residues in the CRD region to penetrate the bilayer, stabilizing complexes through β-strand elongation. These cooperative interactions among lipids, KRAS4b, and RAF1 are proposed as essential for forming nanoclusters, potentially a critical step in MAP kinase signal activation.

A Machine Learning Predictive Model of Bloodstream Infection in Hospitalized Patients.

The aim of the study was to build a machine learning-based predictive model to discriminate between hospitalized patients at low risk and high risk of bloodstream infection (BSI). A Data Mart including all patients hospitalized between January 2016 and December 2019 with suspected BSI was built. Multivariate logistic regression was applied to develop a clinically interpretable machine learning predictive model. The model was trained on 2016-2018 data and tested on 2019 data. A feature selection based on a univariate logistic regression first selected candidate predictors of BSI. A multivariate logistic regression with stepwise feature selection in five-fold cross-validation was applied to express the risk of BSI. A total of 5660 hospitalizations (4026 and 1634 in the training and the validation subsets, respectively) were included. Eleven predictors of BSI were identified. The performance of the model in terms of AUROC was 0.74. Based on the interquartile predicted risk score, 508 (31.1%) patients were defined as being at low risk, 776 (47.5%) at medium risk, and 350 (21.4%) at high risk of BSI. Of them, 14.2% (72/508), 30.8% (239/776), and 64% (224/350) had a BSI, respectively. The performance of the predictive model of BSI is promising. Computational infrastructure and machine learning models can help clinicians identify people at low risk for BSI, ultimately supporting an antibiotic stewardship approach.

The Role of Radical Cystectomy in Clinically Node Positive Bladder Cancer: A US Veterans Health Administration Study

IntroductionThe role of local definitive therapy in addition to systemic treatment in clinically positive regional lymph node (cN+) bladder cancer is yet to be determined. Herein, we sought to investigate the role of radical cystectomy (RC) in management of patients with cN+ bladder cancer at US Veterans Health Administration Facilities. MethodsWe identified patients diagnosed with cN+ bladder cancer between 2000-2017 using the Department of Veterans Affairs (VA) Informatics and Computing Infrastructure (VINCI). We employed a combination of database/registry coded values and chart review for data collection. To minimize mortality bias, we excluded patients who died within 90 days of diagnosis. We divided the patients into cystectomy (C) versus “no cystectomy” (NOC) cohorts. Propensity score matching was performed based on predictors of undergoing RC. Multivariable Cox models and Kaplan-Meier survival curves were used to estimate overall survival (OS) and cancer specific survival (CCS). ResultAfter matching, 158 patients were included in the C and NOC groups. In the C-group, 85(54%) patients received pre-cystectomy chemotherapy, and 73(46%) patients underwent post-cystectomy chemotherapy. In the C-group, 65(41%) patients and in the NOC-group, 66(42%) patients had clinical N1 disease (P = .77). In multivariable Cox model, undergoing RC was associated with improved OS (HR0.62; 95%CI 0.47-0.81), P < .001) and CSS (HR0.58; 95%CI 0.42-0.80; P < .001). ConclusionAs part of multimodal treatment, undergoing RC was associated with improved OS and CSS in subset of patients with cN+ bladder cancer. Prospective randomized trials are warranted to further investigate the role of local definitive therapy in this specific patient population.

Abstract WP249: Risk for Ischemic and Hemorrhagic Stroke is Increased in Veterans Exposed to Agent Orange and Those With Myeloproliferative Neoplasms

Agent Orange (AO) is a dioxin containing defoliant and carcinogen used in the Korean and Vietnam War. There is limited evidence of the association between AO exposure among Veterans and stroke. Stroke is not yet part of the list of presumptive conditions according to the Promise to Address Comprehensive Toxics (PACT) Act which provides Veterans and their survivors disability compensation for conditions arising from exposure to AO. Myeloproliferative Neoplasms (MPN) are uncommon etiologies of stroke but whether AO exposure increases incidence of stroke in MPN has not been described. Utilizing the Veterans Affairs Informatics and Computing Infrastructure (VINCI) database, a case-control study was performed from 1/1/2006 – 1/26/2023 on the Veterans from Illinois, the state most representative of the US population. ICD-9 and -10 codes identified Veterans with stroke and MPN. AO exposure was verified on the Veterans’ service duration and location. Qualitative data were compared by chi-square tests. Among 586,555 Veterans from Illinois, there were 15,455 ischemic stroke (IS), 1,593 hemorrhagic stroke (HS), 2,752 MPN, and 59,393 with AO exposure. Among MPNs, there were 237 IS (41 with AO) and 26 HS (3 with AO). IS and HS were associated with AO exposure, OR 1.34 95% CI 1.28-1.41, p&lt;0.0001, and OR 1.20 95% CI 1.03-1.39, p=0.02, respectively. MPN is associated with IS and HS, OR 3.52, 95% CI 3.08-4.03, and OR 3.54, 95% CI 2.4-5.23, both p&lt;0.0001, respectively. There is no significant association with AO exposure among Veterans with MPN with stroke. Among non-MPN Veterans with AO exposure, there was an earlier median age of IS and HS, 67 vs. 70 and 67 vs. 71, both p&lt;0.0001. There was no difference in median age of stroke among MPN Veterans with or without AO exposure. There were no differences with rates of hypertension, hyperlipidemia, diabetes, smoking, heart failure, and pulmonary hypertension among MPN Veterans with stroke with and without AO exposure. In conclusion, there is an association of AO exposure with IS and HS with an earlier onset among those exposed. There is a strong association between MPN and stroke independent of AO exposure. The biologic plausibility of endothelial dysfunction and accelerated atherosclerosis from AO exposure warrants further investigation.

Achieving High Efficiency: Resource sharing techniques in artificial neural networks for resource-constrained devices

Artificial neural networks emerged in the early 1940s as a connectionist approach to modeling the behavior of interconnected living nerve cells within the intricate cognitive framework of the brain. Over the years, this idea has undergone unprecedented proliferation of ever-expanding application areas, accompanied by an exponential surge in the mathematical formalism employed. This prodigious growth poses severe demands on computational hardware infrastructure. The current research suggests a novel approach for optimizing artificial neural network architectures using resource-sharing techniques. It is discussed the concept of merging context-switching and time-division multiplexing in building a single-neuron artificial neural network with the aid of an FPGA device. This would benefit the significant reduction of hardware resources and the possibility of creating small yet very efficient artificial neural networks without deteriorating the dynamical system performance.

A novel method for effective intrusion detection based on convolutional speaking neural networks

These days, most people can't imagine their life without their networks. People rely more and more on networks as new technologies like 5G and the Internet of Things proliferate, and networks are becoming larger and more complicated by the day. Consequently, the enormous networked society is under increasing danger from cyber attacks, which are also growing in number, sophistication, and variety. Users' private data can end up in the wrong hands. Data communicated via a network might be compromised in some way. It is also possible to target the computer infrastructure that is linked to the network. Thus, NIDS is crucial in ensuring that contemporary society's network communication environment is secure and dependable. This work proposes a detection model that is both lightweight and effective. This model is able to keep its high detection accuracy by combining the benefits of SNN and CNN with logical algorithm design. On top of that, it can bring attention to clear strengths in relation to computing efficiency and power usage. The suggested model was tested against a number of state-of-the-art models using an extensive and integrated set of benchmarks. With a 23 % improvement in detection accuracy (AD), a 19 % improvement in delay, and a 28 % improvement in consumption energy, the suggested technique outperformed the Spiking DFR + MLP and Spike-DHS methods in the simulations.

Abstract TP224: Timing of Ischemic and Hemorrhagic Stroke With Myeloproliferative Neoplasm Diagnosis Among Veterans

Myeloproliferative Neoplasms (MPNs) are rare acquired stem cell disorders, consisting of Essential Thrombocythemia (ET), Polycythemia Vera (PV), and Primary Myelofibrosis (PMF), complicated by thrombohemorrhagic events such as stroke. The timing of ischemic stroke (IS) and hemorrhagic stroke (HS) in relation to MPN diagnosis among Veterans remains undefined. We utilized the Veterans Affairs Informatics and Computing Infrastructure database from 1/1/2006 - 1/26/2023 and included 586,555 Veterans from Illinois, the state most representative of the US population. ICD-9 and -10 codes identified Veterans with stroke and MPN. Fisher’s exact test was utilized to compare MPNs vs. non-MPNs. There were 237 MPNs and 15,221 non-MPNs with IS while there were 26 MPNs and 1,567 non-MPNs with HS (Table 1). There were no differences in age of stroke diagnosis among MPNs vs. non-MPNs. There were higher rates of hypertension, smoking, and heart failure among MPNs vs. non-MPNs with IS. Majority of IS was diagnosed either more than 3 months before MPN diagnosis (N=115, 48.5%) or more than 5 years after diagnosis (N=98, 41.4%). Among MPNs with HS, stroke was predominantly diagnosed (N=14, 53.8%) more than 3 months before and (N=7, 26.9%) within 3 months of MPN diagnosis (Figure 1). Among MPNs there was recurrence of stroke among 45.18% of IS and 40% of HS. Although MPN are infrequent causes of stroke, these findings suggest continued surveillance for MPN and workup for driver mutations after a stroke diagnosis. This study also suggests consideration for dual antiplatelet therapy for MPNs with cardiovascular risk factors to prevent IS and a multidisciplinary approach between neurology and hematology.

Pendampingan Pelatihan Aplikasi Pengolah Kata dan Pengolah Presentasi bagi Siswa SLB Nurul Ikhsan Ngadiluwih

Currently, almost everyone must be able to master technology, especially information technology. The progress of information technology is increasingly rapid, nowadays humans are expected to be able to use computers, laptops, gadgets in everyday life. The aim of this activity is to provide assistance and training in word processing applications (Microsoft Word) and presentation processing (Microsoft Power Point) to Nurul Ikhsan SLB children, because their existence in society is often underestimated. The methods used are introduction, implementation and evaluation. The introduction was carried out by conducting observations and interviews to obtain a general overview of Nurul Ikhsan Ngadiluwih SLB with its various activities. Implementation was carried out by providing training and mentoring to SLB students. After this training the students were very enthusiastic and enthusiastic because our presence was able to provide a different color, their computer training activities were only carried out with limited time in the life skills activity agenda, computer life skills were also very limited because their computer staff and infrastructure were also very limited. There were no obstacles experienced during this service, because they were very welcoming of our presence, and hoped that there would be training on other occasions.

An Architecture and Review of Intelligence Based Traffic Control System for Smart Cities

City traffic congestion can be reduced with the help of adaptable traffic signal control system. The technique improves the efficiency of traffic operations on urban road networks by quickly adjusting the timing of signal values to account for seasonal variations and brief turns in traffic demand. This study looks into how adaptive signal control systems have evolved over time, their technical features, the state of adaptive control research today, and Control solutions for diverse traffic flows composed of linked and autonomous vehicles. This paper finally came to the conclusion that the ability of smart cities to generate vast volumes of information, Artificial Intelligence (AI) approaches that have recently been developed are of interest because they have the power to transform unstructured data into meaningful information to support decision-making (For instance, using current traffic information to adjust traffic lights based on actual traffic circumstances). It will demand a lot of processing power and is not easy to construct these AI applications. Unique computer hardware/technologies are required since some smart city applications require quick responses. In order to achieve the greatest energy savings and QoS, it focuses on the deployment of virtual machines in software-defined data centers. Review of the accuracy vs. latency trade-off for deep learning-based service decisions regarding offloading while providing the best QoS at the edge using compression techniques. During the past, computationally demanding tasks have been handled by cloud computing infrastructures. A promising computer infrastructure is already available and thanks to the new edge computing advancement, which is capable of meeting the needs of tomorrow's smart cities.

Bridging Cognitive Architectures and Generative Models with Vector Symbolic Algebras

Recent developments in generative models have demonstrated that with the right data set, techniques, computational infrastructure, and network architectures, it is possible to generate seemingly intelligent outputs, without explicitly reckoning with underlying cognitive processes. The ability to generate novel, plausible behaviour could be a boon to cognitive modellers. However, insights for cognition are limited, given that generative models' blackbox nature does not provide readily interpretable hypotheses about underlying cognitive mechanisms. On the other hand, cognitive architectures make very strong hypotheses about the nature of cognition, explicitly describing the subjects and processes of reasoning. Unfortunately, the formal framings of cognitive architectures can make it difficult to generate novel or creative outputs. We propose to show that cognitive architectures that rely on certain Vector Symbolic Algebras (VSAs) are, in fact, naturally understood as generative models. We discuss how memories of VSA representations of data form distributions, which are necessary for constructing distributions used in generative models. Finally, we discuss the strengths, challenges, and future directions for this line of work.

Distributed Multi-GPU Ab Initio Density Matrix Renormalization Group Algorithm with Applications to the P-Cluster of Nitrogenase.

The presence of many degenerate d/f orbitals makes polynuclear transition-metal compounds, such as iron-sulfur clusters in nitrogenase, challenging for state-of-the-art quantum chemistry methods. To address this challenge, we present the first distributed multi-graphics processing unit (GPU) ab initio density matrix renormalization group (DMRG) algorithm suitable for modern high-performance computing (HPC) infrastructures. The central idea is to parallelize the most computationally intensive part─the multiplication of O(K2) operators with a trial wave function, where K is the number of spatial orbitals, by combining operator parallelism for distributing the workload with a batched algorithm for performing contractions on GPU. With this new implementation, we are able to reach an unprecedentedly large bond dimension D = 14,000 on 48 GPUs (NVIDIA A100 80 GB SXM) for an active space model (114 electrons in 73 active orbitals) of the P-cluster, which is nearly 3 times larger than the bond dimensions reported in previous DMRG calculations for the same system using only central processing units (CPUs).