Hasil untuk "Computer Science"

Sridhar Chandrasekaran, Yao-Feng Chang, Firman Mangasa Simanjuntak

The memristor has long been known as a nonvolatile memory technology alternative and has recently been explored for neuromorphic computing, owing to its capability to mimic the synaptic plasticity of the human brain. The architecture of a memristor synapse device allows ultra-high-density integration by internetworking with crossbar arrays, which benefits large-scale training and learning using advanced machine-learning algorithms. In this review, we present a statistical analysis of neuromorphic computing device publications from 2018 to 2025, focusing on various memristive systems. Furthermore, we provide a device-level perspective on biomimetic properties in hardware neural networks such as short-term plasticity (STP), long-term plasticity (LTP), spike timing-dependent plasticity (STDP), and spike rate-dependent plasticity (SRDP). Herein, we highlight the utilization of optoelectronic synapses based on 2D materials driven by a sequence of optical stimuli to mimic the plasticity of the human brain, further broadening the scope of memristor controllability by optical stimulation. We also highlight practical applications ranging from MNIST dataset recognition to hardware-based pattern recognition and explore future directions for memristor synapses in healthcare, including artificial cognitive retinal implants, vital organ interfaces, artificial vision systems, and physiological signal anomaly detection.

Erik Arévalo, Ramón Herrera Hernández, Dimitrios Katselis et al.

Direct current motors are widely used in a plethora of applications, ranging from industrial to modern electric (and intelligent) vehicle applications. Most recent operation methods of these motors involve drives that are designed based on an adequate knowledge of the motor dynamics and circulating currents. However, in spite of its simplicity, accurate discrete-time models are not always attainable when utilising the Euler method. Moreover, these inaccuracies may not be reduced when estimating the currents and rotor speed in sensorless direct current motors. In this paper, we analyse three discretisation methods, namely the Euler, second-order Taylor method and second-order Runge–Kutta method, applied to three common types of direct current motor: separately excited, series, and shunt. We also analyse the performance of two of the most simple Bayesian filtering methods, namely the Kalman filter and the extended Kalman filter. For the comparison of the models and the state estimation techniques, we performed several Monte Carlo simulations. Our simulations show that, in general, the Taylor and Runge–Kutta methods exhibit similar behaviours, whilst the Euler method results in less accurate models.

Alejandro Espinosa, Xavier Samos, Daniel Ulied et al.

Abstract As the usage of the edge-cloud continuum increases, Kubernetes presents itself as a solution that allows easy control and deployment of applications in these highly-distributed and heterogeneous environments. In this context, Artificial Intelligence methods have been proposed to aid in the task allocation process to optimize different aspects of the system, such as application execution time, load balancing or energy consumption. In this paper, we present a comparative study focused on optimizing energy consumption through dynamic task allocation in a realistic V2X application scenario. We evaluate and compare two methods representing the most common algorithmic families for resource allocation: Particle Swarm Optimization (PSO) and Proximal Policy Optimization (PPO). Our methodology includes the design of a custom Kubernetes Operator to enforce the models’ node recommendations, allowing for rigorous, real-world validation against the base Kubernetes scheduler. Experiments demonstrate that while both PSO and PPO models successfully reduce energy consumption, PSO delivers the highest savings, reducing energy use by up to 20%. Crucially, our study highlights a key trade-off: although PSO is performance-superior for energy, the PPO model remains a faster and more computationally lightweight option that can be used widely on any kind of device, even with limited resources.

Nathan Tibbetts, Sifat Ibtisum, Satish Puri

The emergence of new, off-path smart network cards (SmartNICs), known generally as Data Processing Units (DPU), has opened a wide range of research opportunities. Of particular interest is the use of these and related devices in tandem with their host's CPU, creating a heterogeneous computing system with new properties and strengths to be explored, capable of accelerating a wide variety of workloads. This survey begins by providing the motivation and relevant background information for this new field, including its origins, a few current hardware offerings, major programming languages and frameworks for using them, and associated challenges. We then review and categorize a number of recent works in the field, covering a wide variety of studies, benchmarks, and application areas, such as data center infrastructure, commercial uses, and AI and ML acceleration. We conclude with a few observations.

Russell Beale

Generative AI tools - most notably large language models (LLMs) like ChatGPT and Codex - are rapidly revolutionizing computer science education. These tools can generate, debug, and explain code, thereby transforming the landscape of programming instruction. This paper examines the profound opportunities that AI offers for enhancing computer science education in general, from coding assistance to fostering innovative pedagogical practices and streamlining assessments. At the same time, it highlights challenges including academic integrity concerns, the risk of over-reliance on AI, and difficulties in verifying originality. We discuss what computer science educators should teach in the AI era, how to best integrate these technologies into curricula, and the best practices for assessing student learning in an environment where AI can generate code, prototypes and user feedback. Finally, we propose a set of policy recommendations designed to harness the potential of generative AI while preserving the integrity and rigour of computer science education. Empirical data and emerging studies are used throughout to support our arguments.

Mizanur Rahman, Amran Bhuiyan, Mohammed Saidul Islam et al.

Recent advances in large language models (LLMs) have enabled a new class of AI agents that automate multiple stages of the data science workflow by integrating planning, tool use, and multimodal reasoning across text, code, tables, and visuals. This survey presents the first comprehensive, lifecycle-aligned taxonomy of data science agents, systematically analyzing and mapping forty-five systems onto the six stages of the end-to-end data science process: business understanding and data acquisition, exploratory analysis and visualization, feature engineering, model building and selection, interpretation and explanation, and deployment and monitoring. In addition to lifecycle coverage, we annotate each agent along five cross-cutting design dimensions: reasoning and planning style, modality integration, tool orchestration depth, learning and alignment methods, and trust, safety, and governance mechanisms. Beyond classification, we provide a critical synthesis of agent capabilities, highlight strengths and limitations at each stage, and review emerging benchmarks and evaluation practices. Our analysis identifies three key trends: most systems emphasize exploratory analysis, visualization, and modeling while neglecting business understanding, deployment, and monitoring; multimodal reasoning and tool orchestration remain unresolved challenges; and over 90% lack explicit trust and safety mechanisms. We conclude by outlining open challenges in alignment stability, explainability, governance, and robust evaluation frameworks, and propose future research directions to guide the development of robust, trustworthy, low-latency, transparent, and broadly accessible data science agents.

Ji-Hoon Kim, Seunghee Han, Kwanghyun Park et al.

The ability to perform machine learning (ML) tasks in a database management system (DBMS) is a new paradigm for conventional database systems as it enables advanced data analytics on top of well-established capabilities of DBMSs. However, the integration of ML in DBMSs introduces new challenges in traditional CPU-based systems because of its higher computational demands and bigger data bandwidth requirements. To address this, hardware acceleration has become even more important in database systems, and the computational storage device (CSD) placing an accelerator near storage is considered as an effective solution due to its high processing power with no extra data movement cost. In this paper, we propose Trinity, an end-to-end database system that enables in-database, in-storage platform that accelerates advanced analytics queries invoking trained ML models along with complex data operations. By designing a full stack from DBMS&#x2019;s internal software components to hardware accelerator, Trinity enables in-database ML pipelines on the CSD. On the software side, we extend the internals of conventional DBMSs to utilize the accelerator in the SmartSSD. Our extended analyzer evaluates the compatibility of the current query with our hardware accelerator and compresses compatible queries into a 24-byte numeric format for efficient hardware processing. Furthermore, the predictor is extended to integrate our performance cost models to always offload queries into the optimal hardware backend. The proposed SmartSSD cost model mathematically models our hardware, including host operations, data transfers, FPGA kernel execution time, and the CPU cost model uses polynomial regression ML models to predict complex CPU latency. On the hardware side, we introduce the in-database processing accelerator (i-DPA), a custom FPGA-based accelerator. i-DPA includes database page decoder to fully exploit the bandwidth benefit of near-storage processing. It also employs dynamic tuple binding to enhance the overall parallelism and hardware utilization. i-DPA;s architecture having heterogeneous computing units with a reconfigurable on-chip interconnect also allows seamless data streaming, enabling task-level pipeline across different computing units. Finally, our evaluation shows that Trinity improves the end-to-end performance of analytics queries by <inline-formula> <tex-math notation="LaTeX">$15.21\times $ </tex-math></inline-formula> on average and up to <inline-formula> <tex-math notation="LaTeX">$57.18\times $ </tex-math></inline-formula> compared to the conventional CPU-based DBMS platform. We also show that the Trinity&#x2019;s performance can linearly scale up with multiple SmartSSDs, achieving nearly up to <inline-formula> <tex-math notation="LaTeX">$200\times $ </tex-math></inline-formula> speedup over the baseline with four SmartSSDs.

Vivekananda Pattanaik, Binaya Kumar Malika, Subhasis Panda et al.

Monitoring, detection, and measurement are vital in the energy system to facilitate better wide-ranging protection, control, and operation. In this regard, to offer a better operational and dynamic performance, Phasor Measurement Units (PMUs) are the prominent components and most desirable choices for transforming the conventional power system into a smart grid and micro-grid-based system. PMUs provide synchronized phasor measurements of electrical parameters such as current, voltage, and other information related to the system's status. However, one of the significant issues related to PMU placement is to place such that the system is fully observable and achieves better state estimation that has a crucial impact on planning, protection, control, and other factors related to the system's overall performance. This motivates the authors to extensively review the innovative approaches suggested by various authors and present their ideas, advantages, limitations, and untouched research gaps. Secondly, this review elaborates on these techniques’ concepts and mathematical details. Thirdly, various techniques are discussed categorically, starting from classical optimization techniques, heuristic and meta-heuristic-based approaches, hybrid techniques, and advanced methods. Also, the problem formulation for the PMUs placement is expressed, giving importance to the objective function, constraints, variables, and assessment index reflecting the optimality in placement. In addition, the future scope is presented to enlighten the researcher on critical discussions regarding the research gaps and fundamental changes to system topology and operation that must be focused on. In this review article, even though the methodologies are similar in approach, a micro-PMU (µPMU) placement is discussed as more important than the PMUs, looking at the present scenario and application in the distribution sector. This article also focuses on various PMU-related standards and real-time applications.

Leah Cathryn Windsor

Child Impact Statements (CIS) are instrumental in helping to foreground the concerns and needs of minor community members who are too young to vote and often unable to advocate for themselves politically. While many politicians and policymakers assert they make decisions in the best interests of children, they often lack the necessary information to meaningfully accomplish this. CISs are akin to Environmental Impact Statements in that both give voice to constituents who are often under-represented in policymaking. This paper highlights an interdisciplinary collaboration between Social Science and Computer Science to create a CIS tool for policymakers and community members in Shelby County, TN. Furthermore, this type of collaboration is fruitful beyond the scope of the CIS tool. Social scientists and computer scientists can leverage their complementary skill sets in data management and data interpretation for the benefit of their communities, advance scientific knowledge, and bridge disciplinary divides within the academy.

Tai Dinh, Wong Hauchi, Daniil Lisik et al.

This paper explores the critical role of data clustering in data science, emphasizing its methodologies, tools, and diverse applications. Traditional techniques, such as partitional and hierarchical clustering, are analyzed alongside advanced approaches such as data stream, density-based, graph-based, and model-based clustering for handling complex structured datasets. The paper highlights key principles underpinning clustering, outlines widely used tools and frameworks, introduces the workflow of clustering in data science, discusses challenges in practical implementation, and examines various applications of clustering. By focusing on these foundations and applications, the discussion underscores clustering's transformative potential. The paper concludes with insights into future research directions, emphasizing clustering's role in driving innovation and enabling data-driven decision-making.

Xianwei ZHU, Wei LIU, Zihao LIU et al.

To address the difficulty faced by network security operation and maintenance personnel in timely and accurate identification of required data during network security event analysis, a recommendation algorithm based on a knowledge graph for network security events was proposed.The algorithm utilized the network threat framework ATT＆amp;CK to construct an ontology model and establish a network threat knowledge graph based on this model.It extracted relevant security data such as attack techniques, vulnerabilities, and defense measures into interconnected security knowledge within the knowledge graph.Entity data was extracted based on the knowledge graph, and entity vectors were obtained using the TransH algorithm.These entity vectors were then used to calculate data similarity between entities in network threat data.Disposal behaviors were extracted from literature on network security event handling and treated as network security data entities.A disposal behavior matrix was constructed, and the behavior matrix enabled the vector representation of network threat data.The similarity of network threat data entities was calculated based on disposal behaviors.Finally, the similarity between network threat data and threat data under network security event handling behavior was fused to generate a data recommendation list for network security events, which established correlations between network threat domains based on user behavior.Experimental results demonstrate that the algorithm performs optimally when the fusion weight α=7 and the recommended data volume K=5, achieving a recall rate of 62.37% and an accuracy rate of 68.23%.By incorporating disposition behavior similarity in addition to data similarity, the algorithm better represents factual disposition behavior.Compared to other algorithms, this algorithm exhibits significant advantages in recall rate and accuracy, particularly when the recommended data volume is less than 10.

Sathya P, Gnanasekaran P

Crop yield forecasting has been well studied in recent decades and is significant in protecting food security. Crop growth is a complex phenomenon that depends on various factors. Machine learning and deep learning trends have emerged as important innovations in this field. We propose to utilize crop, weather, and soil data from agricultural datasets to evaluate yield prediction behavior. Paddy being a staple food crop in India is chosen for this research. In this paper, we propose hybrid architecture for paddy yield prediction, namely, MLR-LSTM, which combines Multiple Linear Regression and Long Short-Term Memory to utilize their complementary nature. The results are compared with traditional machine learning methods such as Support vector machine, Long short-term memory and Random forest method. Evaluation metrics such as Coefficient of Determination (R2), Root Mean Square Error (RMSE), Mean Absolute Error (MAE), Mean Square Error (MSE), F1 score, Recall, and Precision are used to evaluate the hybrid method and traditional models. The results obtained from the proposed hybrid method indicates that the hybrid model delivers better R2, RMSE, MAE, MSE values of 0.93, 0.1549, 0.199, and 0.024 respectively.