This paper presents a lightweight AXI DMA Controller architecture useful for embedded systems that do not require fully featured DMA controllers. Simulation is accomplished with VUnit, and implementation results are o...
详细信息
Amid the rising demand for efficient processors, the challenge has always been to reduce power consumption without compromising performance. FinFET technology has significantly reduced leakage power issues, but dynami...
详细信息
ISBN:
(数字)9798331509118
ISBN:
(纸本)9798331509125
Amid the rising demand for efficient processors, the challenge has always been to reduce power consumption without compromising performance. FinFET technology has significantly reduced leakage power issues, but dynamic power consumption at lower nodes has reemerged as a concern. Methodologies such as power/clock gating, DVFS, and voltage biasing have been proposed in the past. However, next-generation complex SoC designs at < 7nm technology require rethinking at finer granularity. The power gating methodology proposed by You et al., Roy et al., and PG-instr algorithm [1] reduce power consumption but have performance drawbacks. These techniques cause at least a 2% drop in performance in terms of wake-up latency. Moreover, You et al. and Roy et al. only considered idle time, while PG-instr focuses solely on energy savings. None of them balance idle time and energy savings together, which is crucial for achieving the best performance with low power consumption.
computerized test highlights are involved every day in an assortment of ways of understanding biomedical examination, picture control, between PC connections, electronic gadgets and other business exercises. The princ...
详细信息
This paper introduces a first of its kind dataset for audio scene analysis (ASA) and presents a baseline approach for audio source counting. SARdBScene is developed to promote research for audio source counting, as a ...
详细信息
Drones, or unmanned aerial vehicles, have a wide range of uses in a variety of sectors. Surveillance, photography, surveying physically difficult locations, and traffic patrols are some of the applications. License pl...
详细信息
In interpersonal communication, the human face provides important signals of a person's emotional states and intentions. Furthermore, micro-emotions play a major role in understanding hidden intentions. In psychol...
详细信息
The innovative system known as 'Smart Geo-fencing for Religious Places' integrates Blockchain and Zero Power Device technology to enhance the reverence of sacred sites. By amalgamating the robustness of blockc...
详细信息
Lung cancer continues to be the leading cause of cancer-related deaths worldwide, and early detection is essential to extending patient life expectancy. The goal of this study is to boost the precision of lung cancer ...
详细信息
ISBN:
(数字)9798350357509
ISBN:
(纸本)9798350357516
Lung cancer continues to be the leading cause of cancer-related deaths worldwide, and early detection is essential to extending patient life expectancy. The goal of this study is to boost the precision of lung cancer diagnosis through the use of modern machine learning approaches. We developed a thorough data pretreatment strategy that included feature selection, encoding, oversampling, and hyper parameter tuning to handle data discrepancies. We used the Kaggle dataset to assess six ML algorithms: Naive Bayes, Decision Tree, Support Vector Machine (SVM), Random Forest, K-Nearest Neighbors (KNN), and Logistic Regression. With an accuracy of 99.38%, precision of 99.39%, recall of 99.38%, and F1 score of 99.38%, the SVM approach performed exceptionally well.
AI accelerators are ubiquitous across all computational domains, including safety-critical systems such as autonomous driving. However, they are also susceptible to dynamic faults due to aging-related and missed perma...
详细信息
ISBN:
(数字)9798350352597
ISBN:
(纸本)9798350352603
AI accelerators are ubiquitous across all computational domains, including safety-critical systems such as autonomous driving. However, they are also susceptible to dynamic faults due to aging-related and missed permanent manufacturing defects manifesting during in-field operation. Detecting such faults early during in-field operations is essential, particularly for the processing elements with fewer resources. The fault detection framework should operate at the level of predictions rather than individual neurons to achieve fault detection with minimal resource overhead. The output of the classifier network could have captured the fault effects as they propagate through both hardware and software architecture. In this work, we propose a methodology that uses the input image and the logits (output of the classifier network without softmax function) to predict the correctness of the prediction. The method detects the faults that significantly reduce the topl accuracy by more than 50%. This methodology necessitates one-time training of the decoder and additional computation of 1904 neurons using a 3-layer lightweight detection network during in-field operation for fault detection. However, traditional redundancy-based methods such as Dual Modular Redundancy require an additional 4950 neuron computation, equivalent to the classifier network. The proposed method results in around 61 % lesser neuron computation compared to traditional methods and a 99.99 % critical fault detection rate.
Quantum Key Distribution (QKD) is widely accepted as being necessary to counteract threats posed by quantum computers to cybersecurity. However, a QKD setup is fundamentally a single-link solution. Applying this effec...
详细信息
暂无评论