Hasil untuk "Computer engineering. Computer hardware"

Magezi K. Mabaso, Evans M. Chirwa, Shepherd M. Tichapondwa

The high concentration of heavy metals in wastewater highlights the urgent need to explore alternative treatment methods. Partially treated wastewater with elevated heavy metal levels can have severe environmental consequences, ultimately affecting the food chain. This study evaluates the effectiveness of bio-phytoremediation in treating heavy metal-contaminated wastewater using perennial grasses. The research analyzed one-year average effluent results for Pb and Cd, comparing their removal efficiencies at an initial concentration of 10ppm after introducing Vetiver grass (Chrysopogon zizanioides) and Elephant grass (Pennisetum purpurem). The compliance levels of different remediation approaches were assessed against South African wastewater discharge limits and World Health Organization (WHO) guidelines. Various remediation methods were considered, with a particular focus on bio-phytoremediation using selected grass species to remove heavy metals from contaminated wastewater. The findings indicated that Vetiver grass demonstrated a higher removal efficiency for Pb compared to Cd.

Feudjio Ghislain, Saha Tchinda Beaudelaire, Romain Atangana et al.

Context: Early detection of ophthalmic diseases, such as drusen and glaucoma, can be facilitated by analyzing changes in the retinal microvascular structure. The implementation of algorithms based on convolutional neural networks (CNNs) has seen significant growth in the automation of disease identification. However, the complexity of these algorithms increases with the diversity of pathologies to be classified. In this study, we introduce a new lightweight algorithm based on CNNs for the classification of multiple categories of eye diseases, using discrete wavelet transforms to enhance feature extraction. Methods: The proposed approach integrates a simple CNN architecture optimized for multi-class and multi-label classification, with an emphasis on maintaining a compact model size. We improved the feature extraction phase by implementing multi-scale decomposition techniques, such as biorthogonal wavelet transforms, allowing us to capture both fine and coarse features. The developed model was evaluated using a dataset of retinal images categorized into four classes, including a composite class for less common pathologies. Results: The feature extraction based on biorthogonal wavelets enabled our model to achieve perfect values of precision, recall, and F1-score for half of the targeted classes. The overall average accuracy of the model reached 0.9621. Conclusion: The integration of biorthogonal wavelet transforms into our CNN model has proven effective, surpassing the performance of several similar algorithms reported in the literature. This advancement not only enhances the accuracy of real-time diagnoses but also supports the development of sophisticated tools for the detection of a wide range of retinal pathologies, thereby improving clinical decision-making processes.

Kosmas Alexandridis, Vasileios Titopoulos, Giorgos Dimitrakopoulos

Attention mechanisms, particularly within Transformer architectures and large language models (LLMs), have revolutionized sequence modeling in machine learning and artificial intelligence applications. To compute attention for increasingly long sequences, specialized accelerators have been proposed to execute key attention steps directly in hardware. Among the various recently proposed architectures, those based on variants of the FlashAttention algorithm, originally designed for GPUs, stand out due to their optimized computation, tiling capabilities, and reduced memory traffic. In this work, we focus on optimizing the kernel of floating-point-based FlashAttention using new hardware operators that fuse the computation of exponentials and vector multiplications, e.g., e^x, V. The proposed ExpMul hardware operators significantly reduce the area and power costs of FlashAttention-based hardware accelerators. When implemented in a 28nm ASIC technology, they achieve improvements of 28.8% in area and 17.6% in power, on average, compared to state-of-the-art hardware architectures with separate exponentials and vector multiplications hardware operators.

Kathiresan Subramanian, Kagne Suresh

The administration of rural effluent is essential due to the scarcity of resources and infrastructure, which exacerbates health hazards and environmental degradation. Decentralised wastewater treatment (DWT) is a system that reduces the transportation costs and environmental impact by recycling wastewater for agricultural and other purposes. This study suggests a single system DWT technique that integrates natural and artificial systems to enhance rural sustainability. The system employs an Anaerobic Baffled Reactor (ABR) for the initial decomposition of organic matter, a solar-powered disinfection device, and constructed wetlands (CW) for the secondary treatment and nutrient removal. The system is cost-effective and simple to deploy by utilizing renewable energy and local components. The system satisfied local environmental criteria by reducing biochemical oxygen demand (BOD) by 85 %, total suspended solids (TSS) by 90 %, and infections by 99.9 %, as demonstrated by a pilot study conducted in a rural community. Users and administrators were impressed by the system's simplicity and ease of maintenance. Sustainability and ownership were enhanced through community involvement in design and execution. The results indicate that this DWT technique, considered a scalable and adaptive wastewater management solution, has the potential to enhance the quality of water and the health of rural environments. To enhance the sustainability and effectiveness of the system, future research will expand the number of system components and examine alternative applications, such as nitrogen recovery and greywater recycling. This study demonstrates that rural communities may address wastewater treatment challenges by employing a decentralized approach that capitalizes on community engagement and local resources. This strategy also has a positive impact on public health and the environment over time.

Takumi Kobori, Yasunari Suzuki, Yosuke Ueno et al.

Current fault-tolerant quantum computer (FTQC) architectures utilize several encoding techniques to enable reliable logical operations with restricted qubit connectivity. However, such logical operations demand additional memory overhead to ensure fault tolerance. Since the main obstacle to practical quantum computing is the limited qubit count, our primary mission is to design floorplans that can reduce memory overhead without compromising computational capability. Despite extensive efforts to explore FTQC architectures, even the current state-of-the-art floorplan strategy devotes 50% of memory space to this overhead, not to data storage, to ensure unit-time random access to all logical qubits. In this paper, we propose an FTQC architecture based on a novel floorplan strategy, Load/Store Quantum Computer Architecture (LSQCA), which can achieve almost 100% memory density. The idea behind our architecture is to separate all memory regions into small computational space called Computational Registers (CR) and space-efficient memory space called Scan-Access Memory (SAM). We define an instruction set for these abstract structures and provide concrete designs named point-SAM and line-SAM architectures. With this design, we can improve the memory density by allowing variable-latency memory access while concealing the latency with other bottlenecks. We also propose optimization techniques to exploit properties of quantum programs observed in our static analysis, such as access locality in memory reference timestamps. Our numerical results indicate that LSQCA successfully leverages this idea. In a resource-restricted situation, a specific benchmark shows that we can achieve about 90% memory density with 5% increase in the execution time compared to a conventional floorplan, which achieves at most 50% memory density for unit-time random access. Our design ensures broad quantum applicability.

Christodoulos Peltekis, Kosmas Alexandridis, Giorgos Dimitrakopoulos

Transformers have improved drastically the performance of natural language processing (NLP) and computer vision applications. The computation of transformers involves matrix multiplications and non-linear activation functions such as softmax and GELU (Gaussion Error Linear Unit) that are accelerated directly in hardware. Currently, function evaluation is done separately for each function and rarely allows for hardware reuse. To mitigate this problem, in this work, we map the computation of GELU to a softmax operator. In this way, the efficient hardware units designed already for softmax can be reused for computing GELU as well. Computation of GELU can enjoy the inherent vectorized nature of softmax and produce in parallel multiple GELU outcomes. Experimental results show that computing GELU via a pre-existing and incrementally modified softmax hardware unit (a) does not reduce the accuracy of representative NLP applications and (b) allows the reduction of the overall hardware area and power by 6.1% and 11.9%, respectively, on average.

Benchen YANG, Yuhang JIA, Haibo JIN

The overcomplete convolutional structure for biological images and volume segmentation is an excellent solution to the problem in which traditional codec methods cannot accurately segment the boundary regions. Although such methods perform well, the drawback that convolutional operations do not effectively learn global and remote semantic information interactions must be addressed. Accordingly, a new image segmentation network, KTU-Net, is proposed for the medical image segmentation of liver tumors. The network structure constitutes three branches: 1)Kite-Net, which is an overcomplete convolutional network that learns to capture input details and precise edges; 2)U-Net, which learns high-level features; 3)Transformer, which learns sequential representations of input bodies and efficiently captures global multiscale information. KTU-Net is designed for both early and late fusion, and a hybrid loss function is adopted to guide network training to achieve increased stability. From extensive experimental results regarding the LiTS liver tumor segmentation dataset, KTU-Net achieves higher or similar segmentation accuracy than other advanced 3D medical image segmentation methods such as KiU-Net, TransBTS, and UNETR. Fusing the three branching features, the average Dice scores of liver tumors are improved by 0.7% and 2.1%, achieving increased quality of features learned by the network and more accurate segmentation results of liver tumors, thus providing a reliable basis for doctors to perform precise liver tumor cell assessments and treatment plans.

Shafira Putri Ananda, Muhammad Fakhrurrifqi, Divi Galih Prasetyo Putri et al.

Rekam medis adalah catatan kesehatan pasien berdasarkan hasil pemeriksaan, pengobatan, tindakan, dan layanan yang telah diberikan kepada pasien. Beberapa instansi medis telah menggunakan sistem rekam medis elektronik atau Electronic Medical Record (EMR). Namun data rekam medis tersebut hanya dapat diakses secara lokal di tempat fasilitas kesehatan tempat ia berobat sebelumnya. Hal ini mengakibatkan keterbatasan hak akses pasien terhadap rekam medisnya jika ingin berobat ke tempat faskes lain. Berdasarkan permasalahan tersebut, maka dikembangkannya Sistem Informasi DataRawat ini agar pasien dapat mengelola dan membagikan data rekam medisnya kepada pihak yang dikehendaki. Dengan adanya informasi riwayat rekam medis pasien ini akan memudahkan tenaga medis dalam pencarian data informasi rekam medis pasien, mencatat riwayat medis dan mendiagnosa kesehatan pasien dengan cara yang lebih efisien. Sistem informasi ini dikembangkan menggunakan metode Incremental, framework Laravel, bahasa pemrograman PHP dan Javascript, dan PostgreSQL sebagai database. Hasil penelitian ini menunjukkan bahwa Pengembangan Sistem Informasi DataRawat ini dapat mengelola data seperti data rekam medis, data pasien, data rumah sakit, dan membagikan data menggunakan akses link rekam medis. Seluruh fitur yang ada pada analisis kebutuhan dapat terpenuhi dengan baik. Pada pengujian terhadap pengguna hasil menunjukkan skala berkisar 42-48 atau persentase sebesar 84%-96%. Hal ini dapat disimpulkan bahwa sistem masuk ke dalam kategori sangat layak untuk digunakan.

Swati A. Patil, G. Pradeepini, Thirupathi Rao Komati

Abstract Music Genre Classification (MGC) is a crucial undertaking that categorizes Music Genre (MG) based on auditory information. MGC is commonly employed in the retrieval of music information. The three main stages of the proposed system are data readiness, feature mining, and categorization. To categorize MG, a new neural network was deployed. The proposed system uses features from spectrographs derived from short clips of songs as inputs to a projected scheme building to categorize songs into an appropriate MG. Extensive experiment on the GTZAN dataset, Indian Music Genre(IMG) dataset, Hindustan Music Rhythm (HMR) and Tabala Dataset show that the proposed strategy is more effective than existing methods. Indian rhythms were used to test the proposed system design. The proposed system design was compared with other existing algorithms based on time and space complexity.

SHEN Xueli, HAN Qianwen

Click-Through Rate(CTR) prediction is one of the most important tools for ad placement.Predicting the CTR of an ad and making recommendations to users can increase ad revenue.Field-aware click-through rate prediction models are superior to other click-through rate prediction models because they consider the field information; however, they generate a large amount of redundant information during feature interaction, which results in a low prediction accuracy.A Field-aware Attention Embedding Neural Network(FAENN) model is herein proposed.This model uses a Self-Attentive Mechanism(SAM) to distribute weights to the input vectors of the embedding layer.This helps to clearly identify the importance of the field-aware embedded features, speeding up the training process.The lower-order feature interaction layer focuses on the explicit first-order information of the features and the second-order interaction feature information and outputs the effective features to the higher-order interaction layer.The higher-order feature interaction layer combines the learned interaction vectors with the deep neural network to capture higher-order feature interactions to improve prediction accuracy.The experimental results show that the FAENN model has a higher prediction accuracy than the FM, FFM, and AFM models.

Milin Zhang, Mohammad Abdi, Jonathan Ashdown et al.

Distributed deep neural networks (DNNs) have been shown to reduce the computational burden of mobile devices and decrease the end-to-end inference latency in edge computing scenarios. While distributed DNNs have been studied, to the best of our knowledge, the resilience of distributed DNNs to adversarial action remains an open problem. In this paper, we fill the existing research gap by rigorously analyzing the robustness of distributed DNNs against adversarial action. We cast this problem in the context of information theory and rigorously proved that (i) the compressed latent dimension improves the robustness but also affect task-oriented performance; and (ii) the deeper splitting point enhances the robustness but also increases the computational burden. These two trade-offs provide a novel perspective to design robust distributed DNN. To test our theoretical findings, we perform extensive experimental analysis by considering 6 different DNN architectures, 6 different approaches for distributed DNN and 10 different adversarial attacks using the ImageNet-1K dataset.

Md Sakib Hasan, Catherine D. Schuman, Zhongyang Zhang et al.

Neuromorphic Computing promises orders of magnitude improvement in energy efficiency compared to traditional von Neumann computing paradigm. The goal is to develop an adaptive, fault-tolerant, low-footprint, fast, low-energy intelligent system by learning and emulating brain functionality which can be realized through innovation in different abstraction layers including material, device, circuit, architecture and algorithm. As the energy consumption in complex vision tasks keep increasing exponentially due to larger data set and resource-constrained edge devices become increasingly ubiquitous, spike-based neuromorphic computing approaches can be viable alternative to deep convolutional neural network that is dominating the vision field today. In this book chapter, we introduce neuromorphic computing, outline a few representative examples from different layers of the design stack (devices, circuits and algorithms) and conclude with a few exciting applications and future research directions that seem promising for computer vision in the near future.

DIVYA, T., RAMAPRABHA, R.

The modeling and the implementation of the embedded switched z-source type cascaded multilevel inverter for photovoltaic (PV) interfaced applications have been presented. With the ability to draw continuous current with an inherent filtering capability, the embedded switched z-source type inverter provides a single-stage conversion with a high output gain which makes it suitable for PV arrays with a low voltage rating. By applying a modular cascading method with a reduced number of H-bridge the multilevel inverter (MLI) is designed for a series-parallel connected PV array. It is controlled using the basic multicarrier PWM technique and synchronized with the grid. With the derived design equations for each mode, its stability has been analyzed and compared for different duty cycles. The developed MLI connected with a PV array has been simulated with the idea of reducing the impact of partial shading by using shorter series of strings, providing a high gain conversion with lower stress across the components. A prototype of the MLI has been tested to give a power rating of 2 kW and the results from both the simulation and the hardware have been discussed.

Nathalia Raquel Salazar Tigrero

El objetivo principal de este artículo se enfoca en el análisis del principio de estabilidad laboral de las personas con enfermedades catastróficas, como un derecho fundamental que permite el goce de una protección diferenciada en el ámbito laboral, en vista de su condición de doble vulnerabilidad reconocida en la Constitución y en la ley Orgánica de la Salud. La metodología científica ha permitido revelar la gran importancia de este principio en la protección de los derechos de los trabajadores. La revisión bibliográfica de la literatura se hizo utilizando las Tecnologías de la Información y las Comunicaciones (TIC) para contextualizar esta área de particular interés a través de la identificación, evaluación e interpretación del conjunto de trabajos de investigación que describen dicha área. Los resultados obtenidos reflejaron que lamentablemente, la inefectiva aplicación de este principio en el Ecuador ha generado frecuentemente condiciones no favorables para estos y que se traducen en la vulneración de sus derechos. Se hace importante que las instituciones gubernamentales produzcan nuevas políticas públicas que garanticen la honra y dignidad de estas personas que llevan una doble condición de vulnerabilidad en el Ecuador.

Yu-Hsiang Chiang, Cheng En Ni, Yun Sung et al.

In-memory computing is becoming a popular architecture for deep-learning hardware accelerators recently due to its highly parallel computing, low power, and low area cost. However, in-RRAM computing (IRC) suffered from large device variation and numerous nonideal effects in hardware. Although previous approaches including these effects in model training successfully improved variation tolerance, they only considered part of the nonideal effects and relatively simple classification tasks. This paper proposes a joint hardware and software optimization strategy to design a hardware-robust IRC macro for object detection. We lower the cell current by using a low word-line voltage to enable a complete convolution calculation in one operation that minimizes the impact of nonlinear addition. We also implement ternary weight mapping and remove batch normalization for better tolerance against device variation, sense amplifier variation, and IR drop problem. An extra bias is included to overcome the limitation of the current sensing range. The proposed approach has been successfully applied to a complex object detection task with only 3.85\% mAP drop, whereas a naive design suffers catastrophic failure under these nonideal effects.

Omar Costa Hamido

In this chapter I will discuss the role of quantum computing in computer music and how it can be integrated to better serve the creative artists. I will start by considering different approaches in current computer music and quantum computing tools, as well as reviewing some previous attempts to integrate them. Then, I will reflect on the meaning of this integration and present what I coined as QAC (Quantum-computing Aided Composition) as well as an early attempt at realizing it. This chapter will also introduce The QAC Toolkit Max package, analyze its performance, and explore some examples of what it can offer to realtime creative practice. Lastly, I will present a real case scenario of QAC in the creative work Disklavier Prelude #3.

Bhaskar DasGupta, Elena Grigorescu, Tamalika Mukherjee

Characterizing shapes of high-dimensional objects via Ricci curvatures plays a critical role in many research areas in mathematics and physics. However, even though several discretizations of Ricci curvatures for discrete combinatorial objects such as networks have been proposed and studied by mathematicians, the computational complexity aspects of these discretizations have escaped the attention of theoretical computer scientists to a large extent. In this paper, we study one such discretization, namely the Ollivier-Ricci curvature, from the perspective of efficient computation by fine-grained reductions and local query-based algorithms. Our main contributions are the following. (a) We relate our curvature computation problem to minimum weight perfect matching problem on complete bipartite graphs via fine-grained reduction. (b) We formalize the computational aspects of the curvature computation problems in suitable frameworks so that they can be studied by researchers in local algorithms. (c) We provide the first known lower and upper bounds on queries for query-based algorithms for the curvature computation problems in our local algorithms framework. En route, we also illustrate a localized version of our fine-grained reduction. We believe that our results bring forth an intriguing set of research questions, motivated both in theory and practice, regarding designing efficient algorithms for curvatures of objects.

Andriy Miranskyy, Mushahid Khan, Jean Paul Latyr Faye et al.

Quantum computers (QCs) are maturing. When QCs are powerful enough, they may be able to handle problems in chemistry, physics, and finance that are not classically solvable. However, the applicability of quantum algorithms to speed up Software Engineering (SE) tasks has not been explored. We examine eight groups of quantum algorithms that may accelerate SE tasks across the different phases of SE and sketch potential opportunities and challenges.

Evan McKinney, Mingkang Xia, Chao Zhou et al.

Noisy, Intermediate Scale Quantum (NISQ) computers have reached the point where they can show the potential for quantum advantage over classical computing. Unfortunately, NISQ machines introduce sufficient noise that even for moderate size quantum circuits the results can be unreliable. We propose a co-designed superconducting quantum computer using a Superconducting Nonlinear Asymmetric Inductive eLement (SNAIL) modulator. The SNAIL modulator is designed by considering both the ideal fundamental qubit gate operation while maximizing the qubit coupling capabilities. First, the SNAIL natively implements $\sqrt[n]{\texttt{iSWAP}}$ gates realized through proportionally scaled pulse lengths. This naturally includes $\sqrt{\texttt{iSWAP}}$, which provides an advantage over $\texttt{CNOT}$ as a basis gate. Second, the SNAIL enables high-degree couplings that allow rich and highly parallel qubit connection topologies without suffering from frequency crowding. Building on our previously demonstrated SNAIL-based quantum state router we propose a quantum 4-ary tree and a hypercube inspired corral built from interconnected quantum modules. We compare their advantage in data movement based on necessary \texttt{SWAP} gates to the traditional lattice and heavy-hex lattice used in latest commercial quantum computers. We demonstrate the co-design advantage of our SNAIL-based machine with $\sqrt{\texttt{iSWAP}}$ basis gates and rich topologies against $\texttt{CNOT}$/heavy-hex and $\texttt{FSIM}$/lattice for 16-20 qubit and extrapolated designs circa 80 qubit architectures. We compare total circuit time and total gate count to understand fidelity for systems dominated by decoherence and control imperfections, respectively. Finally, we provide a gate duration sensitivity study on further decreasing the SNAIL pulse length to realize $\sqrt[n]{\texttt{iSWAP}}$ qubit systems to reduce decoherence times.

Boyu Ni, Xiaoyang Dong, Keting Jia et al.

Simpira v2 is an AES-based permutation proposed by Gueron and Mouha at ASIACRYPT 2016. In this paper, we build an improved MILP model to count the differential and linear active Sboxes for Simpira v2, which achieves tighter bounds of the minimum number of active Sboxes for a few versions of Simpira v2. Then, based on the new model, we find some new truncated differentials for Simpira v2 and give a series (quantum) collision attacks on two versions of reduced Simpira v2.