Explanation-guided learning (EGL) has gained prominence for improving both the explainability and performance of deep neural networks by integrating additional supervision signals based on the elucidation of the model...
详细信息
A wearable microgrid that centralizes and distributes harvested energy across different body regions can optimize power utilization and reduce overall battery *** setup underscores the importance of developing cable-f...
详细信息
A wearable microgrid that centralizes and distributes harvested energy across different body regions can optimize power utilization and reduce overall battery *** setup underscores the importance of developing cable-free wireless power transfer(WPT)systems for mobile and portable devices to eliminate the risks posed by wired connections,especially in dynamic and hazardous *** introduce a thin,stretchable,and safe hand band capable of watt-level wireless charging through the widely adopted Qi protocol operating at 130 *** implementation of nonadhesive fabric encapsulation serves to protect the 50-μm-thin spiral copper antenna from mechanical strain,ensuring an overall hand band stretchability of 50%.We also create a stretchable“Ferrofabric”,characterized by a magnetic permeability of 11.3 and a tensile modulus of 75.3 kPa,that provides magnetic shielding for the antenna without compromising ***“Ferrofabric”improves the coil inductance but induces core loss in AC *** fully understanding and managing loss mechanisms such as the skin effect,proximity effect,core loss,and joule heating,we achieve a wireless charging efficiency of71%andpower delivery of 3.81Win the kHz frequency *** band is unobstructive to hand motion and can charge a handheld smartphone as fast as a desktop charger or power a battery-free chest-laminated e-tattoo sensor,with well-managed thermal and electromagnetic *** a holistic electromagnetic,structural,and thermal design,our device culminates in a safe,rugged,and versatile solution for wearable WPT systems.
Accurately predicting the Remaining Useful Life(RUL)of lithium-ion batteries is crucial for battery management *** learning-based methods have been shown to be effective in predicting RUL by leveraging battery capacit...
详细信息
Accurately predicting the Remaining Useful Life(RUL)of lithium-ion batteries is crucial for battery management *** learning-based methods have been shown to be effective in predicting RUL by leveraging battery capacity time series ***,the representation learning of features such as long-distance sequence dependencies and mutations in capacity time series still needs to be *** address this challenge,this paper proposes a novel deep learning model,the MLP-Mixer and Mixture of Expert(MMMe)model,for RUL *** MMMe model leverages the Gated Recurrent Unit and Multi-Head Attention mechanism to encode the sequential data of battery capacity to capture the temporal features and a re-zero MLP-Mixer model to capture the high-level ***,we devise an ensemble predictor based on a Mixture-of-Experts(MoE)architecture to generate reliable RUL *** experimental results on public datasets demonstrate that our proposed model significantly outperforms other existing methods,providing more reliable and precise RUL predictions while also accurately tracking the capacity degradation *** code and dataset are available at the website of github.
CsSnI3 is widely studied as an environmentally friendly Pb-free perovskite material for optoelectronic device applications. To further improve material and device performance, it is important to understand the surface...
详细信息
CsSnI3 is widely studied as an environmentally friendly Pb-free perovskite material for optoelectronic device applications. To further improve material and device performance, it is important to understand the surface structures of CsSnI3. We generate surface structures with various stoichiometries, perform density functional theory calculations to create phase diagrams of the CsSnI3 (001), (110), and (100) surfaces, and determine the most stable surfaces under a wide range of Cs, Sn, and I chemical potentials. Under I-rich conditions, surfaces with Cs vacancies are stable, which lead to partially occupied surface states above the valence band maximum. Under I-poor conditions, we find the stoichiometric (100) surface to be stable under a wide region of the phase diagram, which does not have any surface states and can contribute to long charge-carrier lifetimes. Consequently, the I-poor (Sn-rich) conditions will be more beneficial to improve the device performance.
This study investigates the biological effects of Low-Frequency Electromagnetic Fields (LF-EMF) on Saccharomyces cerevisiae yeast cells, an essential model organism in biological research. To achieve controlled exposu...
详细信息
Accurate prediction of the remaining useful life(RUL)is crucial for the design and management of lithium-ion *** various machine learning models offer promising predictions,one critical but often overlooked challenge ...
详细信息
Accurate prediction of the remaining useful life(RUL)is crucial for the design and management of lithium-ion *** various machine learning models offer promising predictions,one critical but often overlooked challenge is their demand for considerable run-to-failure data for *** of such training data leads to prohibitive testing efforts as the run-to-failure tests can last for ***,we propose a semi-supervised representation learning method to enhance prediction accuracy by learning from data without RUL *** approach builds on a sophisticated deep neural network that comprises an encoder and three decoder heads to extract time-dependent representation features from short-term battery operating data regardless of the existence of RUL *** approach is validated using three datasets collected from 34 batteries operating under various conditions,encompassing over 19,900 charge and discharge *** method achieves a root mean squared error(RMSE)within 25 cycles,even when only 1/50 of the training dataset is labelled,representing a reduction of 48%compared to the conventional *** also demonstrate the method's robustness with varying numbers of labelled data and different weights assigned to the three decoder *** projection of extracted features in low space reveals that our method effectively learns degradation features from unlabelled *** approach highlights the promise of utilising semi-supervised learning to reduce the data demand for reliability monitoring of energy devices.
The digital transformation process of power systems towards smart grids is resulting in improved reliability, efficiency and situational awareness at the expense of increased cybersecurity vulnerabilities. Given the a...
详细信息
The digital transformation process of power systems towards smart grids is resulting in improved reliability, efficiency and situational awareness at the expense of increased cybersecurity vulnerabilities. Given the availability of large volumes of smart grid data, machine learning-based methods are considered an effective way to improve cybersecurity posture. Despite the unquestionable merits of machine learning approaches for cybersecurity enhancement, they represent a component of the cyberattack surface that is vulnerable, in particular, to adversarial attacks. In this paper, we examine the robustness of autoencoder-based cyberattack detection systems in smart grids to adversarial attacks. A novel iterative-based method is first proposed to craft adversarial attack samples. Then, it is demonstrated that an attacker with white-box access to the autoencoder-based cyberattack detection systems can successfully craft evasive samples using the proposed method. The results indicate that naive initial adversarial seeds cannot be employed to craft successful adversarial attacks shedding insight on the complexity of designing adversarial attacks against autoencoder-based cyberattack detection systems in smart grids.
The impact of orthopedic scaffolds on bone defect healing,particularly the late-stage bone remodeling process,is pivotal for the therapeutic *** study applies fadditively manufactured scaffolds composed of hydroxyapat...
详细信息
The impact of orthopedic scaffolds on bone defect healing,particularly the late-stage bone remodeling process,is pivotal for the therapeutic *** study applies fadditively manufactured scaffolds composed of hydroxyapatite-doped poly(lactide-co-glycolide)-b-poly(ethylene glycol)-b-poly(lactide-co-glycolide)(HAPELGA)with varying properties to treat rat calvarial defects,elucidating their significant role in bone remodeling by modulating physiological *** engineered two scaffolds with different polylactic acid(PLA)to polyglycolic acid(PGA)ratio(9/1 and 18/1)to vary in hydrophobicity,degradation rate,mechanical properties,and structural *** variations influenced physiological responses,including osteogenesis,angiogen-esis,and immune reactions,thereby guiding bone *** findings show that the HA-PELGA(18/1)scaffold,with a slower degradation rate,supported bulk bone formation due to a stable ***,the HA-PELGA(9/1)scaffold,with a faster degradation rate and more active interfaces,facilitated the formation of a thin bone layer and higher bone *** study demonstrates these degradable scaffolds help to promote bone healing and reveals how scaffold properties influence the bone remodeling process,offering a potential strategy to optimize scaffold design aiming at late-stage bone defect healing.
Deep learning has been proved to diagnose Attention Deficit/Hyperactivity Disorder (ADHD) accurately, but it has raised concerns about trustworthiness because of the lack of explainability. Fortunately, the developmen...
详细信息
This study presents a performance comparison of different deep machine learning models for the intelligent segregation of date fruit bunches of numerous varieties. The shape, size, and color of the various types of da...
详细信息
暂无评论