21,089 research outputs found
Low-Complexity On-Device ECG Classifier using Binarized Neural Network
No full textTargeting the real-Time arrhythmia diagnosis on resource-limited devices, in this paper, we present a cost-effective heartbeat classifier that uses a binarized neural network (BNN). Based on the previous CNN-based approaches, several optimization schemes are applied for error-resilient binarization. With the full-precision gradient descent, the proposed model reduces the variance during the training. The learnable activation function additionally compensates binarization errors by adjusting information shifts. Targeting the MIT-BIH arrhythmia database, allowing less than 1% accuracy drop, the proposed BNN model reduces the energy consumption and the memory usage by 94.63% and 84.62%, respectively, compared to the full-precision CNN-based model.2
Performance analysis of a hybrid HVAC system consisting of a solar thermal collector and a radiative cooling panel
No full textWe proposed a hybrid heating, ventilation, and air-conditioning (HVAC) system that integrates a solar thermal collector and a radiative cooling panel as a heat source and a heat sink, respectively. The annual performance of the proposed system was analyzed for a single-family house and compared with that of the conventional solar-assisted heat pump (SAHP) and radiative-cooling-assited heat pump (RAHP) sys-tems. We found that the seasonal coefficient of performance of the proposed system can be enhanced by 27.3% for cooling compared to the SAHP system in cold-dry climate and by 61.8% for heating compared to the RAHP system in hot-dry climate. The detailed analysis revealed that the sizes of the solar collector and the radiative cooling panel, and the total amount of water (i.e., working as a heat source or a heat sink for the proposed HVAC system) are the key parameters affecting the annual performance of the pro-posed HAHP system. By using an optimization technique, we showed that these parameters should be varied depending on climate zones for optimal operation. Interestingly, it was found that installing both solar collector and radiative cooing panel can be more efficient than the conventional systems regardless of climate conditions (i.e., cold-dry: Denver; hot-dry: Pheonix; and marine: Los Angeles). The sensitivity analysis of each parameter revealed that the annual power consumption is less sensitive to the tank water volume for cold-dry and hot-dry climates, but the tank volume becomes important for marine cli-mate. For considered three locations, the proposed system was found to be able to reduce the annual power consumption by about 3% ti 29% as compared to the conventional RAHP or SAHP system. (c) 2021 Elsevier B.V. All rights reserved.
‐maleimide)‐Incorporated Poly(vinylidene fluoride)
No full textA novel electro-active polymer actuator employing the ionic networking membrane of poly(styrene-alt-maleimide) (PSMI)-incorporated poly(vinylidene fluoride) (PVDF) was developed to improve the electrical and mechanical performance of the artificial muscles. The main drawback of the previous ionic polymer-metal composite actuator was the straightening-back and relaxation under the constant voltage excitation. The present ionic networking membrane actuator overcomes the relaxation of the ionic polymer-metal composite actuator under the constant voltage and also shows much larger tip displacement than that of the Nafion-based actuator. Under the simple harmonic stimulus, the measured mechanical displacement was comparable to that of the Nafion-based actuator. The excellent electromechanical response of the current polymer actuator is attributed to two factors: the inherent large ionic-exchange capacity and the unique hydrophilic nano-channels of the ionic networking membrane. The electro-active polymer actuator of PSMI-incorporated PVDF can be a promising smart material and may possibly diversify niche applications in biomimetic motion
A Lightweight ML-Based ECG Classification System Using Self-Personalized Anomaly Detector
No full textTargeting the real-time arrhythmia diagnosis on resource-limited edge devices, in this paper, we present a lightweight electrocardiogram classification system using event-driven machine learning processing. A self-personalized anomaly detector based on signal processing is newly developed to dynamically update internal decision criteria from each patient's recent electrocardiogram history, that activates the following machine learning model only for the abnormal cases. A Siamese neural network is adopted to identify detailed arrhythmia classes by comparing features from the self-personalized normal data and the current abnormal input, increasing the classification accuracy. We also develop a simple version of our Siamese model to reduce the number of trainable parameters while preserving the end-to-end classification accuracy. Experimental results show that the proposed event-driven system reduces ML model activations by 74% for normal beats, achieving a classification accuracy of 96.9% comparable to leading solutions. Additionally, it consumes three times less energy and achieves 3.6 times faster processing latency compared to cost-aware method on a mobile GPU platform, enabling extended battery life and real-time analysis on edge devices.
Novel electro-active polymer actuator based on ionic networking membrane of PSMI-incorporated PVDF
No full textFabrication and actuation of electro-active polymer actuator based on PSMI-incorporated PVDF
No full textIn this study, an ionic networking membrane (INM) of poly(styrene-alt-maleimide) (PSMI)-incorporated poly(vinylidene fluoride) (PVDF) was applied to fabricate electro-active polymer. Based on the same original membrane of PSMI-incorporated PVDF, various samples of INM actuator were prepared for different reduction times with the electroless-plating technique. The as-prepared INM actuators were tested in terms of surface resistance, platinum morphology, resonance frequency, tip displacement, current and blocked force, and their performances were compared to those of the widely used traditional Nafion actuator. Scanning electron microscopy (SEM) and transmission electron microscopy (TEM) revealed that much smaller and more uniform platinum particles were formed on the surfaces of the INM actuators as well as within their polymer matrix. Although excellent harmonic responses were observed for the newly developed INM actuators, they were found to be sensitive to the applied reduction times during the fabrication. The mechanical displacement of the INM actuator fabricated after the optimum reduction times was much larger than that of its Nafion counterpart of comparable thickness under the stimulus of constant and alternating current voltage. The PSMI-incorporated PVDF actuator can become a promising smart material to be used in the fields of biomimetic robots, biomedical devices, sensors and actuator, haptic interfaces, energy harvesting and so on
A Review of the Theory and a Case in Korean Conservatism: on Sunwoo His Perspective
No full text그간 많은 선행 연구들이 한국 보수주의의 사상적 빈곤성을 지적해 왔지만 존재론적으로도 한국에 보수주의적 가치가 전무했다고 단정할 수는 없다. 평가적 가치 판단에 앞서, 한국 보수주의의 기원과 실체를 규명하는 작업이 필수적이다. 세 가지 작업을 수행한다. 첫째, 근대 보수주의가 갖는 존재론적 속성과 일반적 구성 요소를 검토한다. 우리에 알려진 보수주의가 서구 역사를 기초로 형성된 맥락을 유념하여, 선행 연구들이 정리한 보수주의의 구성 요소들 중 서구적 특수성을 소거하는 방식으로 근대 보수주의의 일반적 구성 요소를 재구성한다. 둘째, 한국 보수주의의 이론적 특성 규명이다. 보수주의의 일반적 구성 요소들은 한국 현대정치의 특수한 조건 및 그에 기인한 가치들과 접목되어 한국 보수주의를 형성해 왔다. 한국 현대정치를 규정한 주요한 조건인 반공주의, 발전국가, 유교 전통 세 요소가 유기적으로 결합하여 한국 나름의 보수주의적 가치의 기초가 되어 왔다. 셋째, 한국 보수주의의 사례 분석이다. 20세기 후반 대표적 보수주의자로 꼽히는 언론인 선우휘의 입장에 이러한 보수적 가치들이 어떻게 용해되었는지 그의 논설과 소설을 인용하며 검토한다.The purpose of this paper is to analyze the theory and thoughts of conservatism in modern Korean politics and Sunwoo His case as a representative conservative intellectual in 20th century. According to this paper, I argue that the basic conservative virtues and principles formed in the period of military government from 1960s to 1980s, when it comes to the contemporary Korean politics. And I analyze the general characters of modern conservatism and Koreas unique experience in the context of their historical course. The difference between Western conservatism and Korean conservatism due to Koreas modern political history. Sunwoo His conservatism presents several unique elements and general characters. Political elitism, simple life of individual, and strong anti-Communism, statism. From this point of view, this analyze of theory and case indicate the origin and causes of todays Korean conservative force and conservatism
Multi-behavior Recommender Systems: A Survey
No full textTraditional recommender systems primarily rely on a single type of user-item interaction, such as item purchases or ratings, to predict user preferences. However, in real-world scenarios, users engage in a variety of behaviors, such as clicking on items or adding them to carts, offering richer insights into their interests. Multi-behavior recommender systems leverage these diverse interactions to enhance recommendation quality, and research on this topic has grown rapidly in recent years. This survey provides a timely review of multi-behavior recommender systems, focusing on three key steps: (1) Data Modeling: representing multi-behaviors at the input level, (2) Encoding: transforming these inputs into vector representations (i.e., embeddings), and (3) Training: optimizing machine-learning models. We systematically categorize existing multi-behavior recommender systems based on the commonalities and differences in their approaches across the above steps. Additionally, we discuss promising future directions for advancing multi-behavior recommender systems
- …
