A great number of deep learning-based models have been recently proposed for automatic piano classification. In this paper, we describe our contribution to the challenge of automatic piano classification when the perf...
详细信息
ISBN:
(数字)9798350386844
ISBN:
(纸本)9798350386851
A great number of deep learning-based models have been recently proposed for automatic piano classification. In this paper, we describe our contribution to the challenge of automatic piano classification when the performer performs at the concert or stage. Among these models in deep learning, we use init-1D-WaveNet and init-2D-MLNet for comparison the accuracy in the piano beginning level of the Christmas song (Jingle bells). Our experimental results show that the assessment using the init-2D-MLNet still achieve high accuracy of 87.5%.
作者：
Shi, ChanghaoMishne, GalUC San Diego
Electrical and Computer Engineering Department CA92093 United States UC San Diego
Hal.c.o.glu Data Science Institute and the Neurosciences Graduate Program CA92093 United States
Graph signal processing (GSP) is a prominent framework for analyzing signals on non-Euclidean domains. The graph Fourier transform (GFT) uses the combinatorial graph Laplacian matrix to reveal the spectral decompositi...
详细信息
This research-to-practice paper describes developing and analyzing state-of-the-art smart boots created by combining CAD technology and advanced 3D printing techniques to attract students in bio-engineering and relate...
详细信息
ISBN:
(数字)9798350351507
ISBN:
(纸本)9798350363067
This research-to-practice paper describes developing and analyzing state-of-the-art smart boots created by combining CAD technology and advanced 3D printing techniques to attract students in bio-engineering and related fields. The primary objective of this innovative immobilization boot is to expedite fracture recovery phases through an ergonomic design to ensure optimal patient comfort during its use. Technological solutions are crucial in aiding the rehabilitation process for fractures caused by falls, heavy lifting, or rotational trauma. However, cost and comfort-related issues persist, underscoring the need for alternative approaches. This research addresses these challenges and delves into the broader implications of fracture treatment, catalyzing future projects and investigations in bioengineering. Additionally, this study serves as an educational tool that sparks the interest of high school and engineering students, promoting multidisciplinary collaboration in innovation. By involving students in specialized courses covering 3D design, human bone anatomy, biology, and materials science, this initiative empowers them to deepen their knowledge and develop new technologies to address bone injury problems. Material analyses include evaluating the type of material depending on the fracture site, such as PLA for printing and cotton and silicone gel for the midsection between the splint and the body. This research aims to advance our understanding of the type of fracture, the methods associated with their treatment, and tissue repair processes during bone callus formation. To summarize, this multidisciplinary approach drives advancements in bio-engineering and related fields, aiming to enhance patient outcomes and inspire students to pursue further research in bio-engineering and related fields. As part of this endeavor, a list of university-level courses based on the experience of the University of Puerto Rico at Mayaguez (UPRM), such as biology, bio-materials, 3D
Impedance-based protein detection sensors for point-of-care diagnostics require quantitative specificity,as well as rapid or real-time ***,microfabrication of these sensors can lead to the formation of factors suitabl...
详细信息
Impedance-based protein detection sensors for point-of-care diagnostics require quantitative specificity,as well as rapid or real-time ***,microfabrication of these sensors can lead to the formation of factors suitable for in vivo ***,we present microfabricated needle-shaped microwell impedance sensors for rapid-sample-to-answer,label-free detection of cytokines,and other *** microneedle form factor allows sensors to be utilized in transcutaneous or transvascular sensing *** vitro,experimental characterization confirmed sensor specificity and sensitivity to multiple proteins of *** characterization demonstrated sufficient microneedle robustness for transcutaneous insertion,as well as preserved sensor function *** further utilized these sensors to carry out real-time in vivo quantification of human interleukin 8(hIL8)concentration levels in the blood of transgenic mice that endogenously express *** assess sensor functionality,hIL8 concentration levels in serum samples from the same mice were quantified by *** agreement between real-time in vivo sensor readings in blood and subsequent ELISA serum assays was observed over multiple transgenic mice expressing hIL8 concentrations from 62 pg/mL to 539 ng/mL.
The Neural Networks (NN) model which is incorporated in the control system design has been studied, and the results show better performance than the mathematical model approach. However, some studies consider that onl...
详细信息
The Neural Networks (NN) model which is incorporated in the control system design has been studied, and the results show better performance than the mathematical model approach. However, some studies consider that only offline NN model learning and does not use the online NN model learning directly on the control system. As a result, the controller's performance decreases due to changes in the system environment from time to time. The Reinforcement Learning (RL) method has been investigated intensively, especially Model-based RL (Mb-RL) to predict system dynamics. It has been investigated and performs well in making the system more robust to environmental changes by enabling online learning. This paper proposes online learning of local dynamics using the Mb-RL method by utilizing Long Short-Term Memory (LSTM) model. We consider Model Predictive Control (MPC) scheme as an agent of the Mb-RL method to control the regulatory trajectory objectives with a random shooting policy to search for the minimum objective function. A nonlinear Mass Spring Damper (NMSD) system with parameter-varying linear inertia is used to demonstrate the effectiveness of the proposed method. The simulation results show that the system can effectively control high-oscillating nonlinear systems with good performance.
A strategy that combines experiment and simulation to design and optimize electromagnetic (EM) metamaterial absorbers containing a periodic porous structure is described. The approach provides the ability to produce a...
A strategy that combines experiment and simulation to design and optimize electromagnetic (EM) metamaterial absorbers containing a periodic porous structure is described. The approach provides the ability to produce absorbers that meet multiple user-specified objectives. Using the measured intrinsic properties of the baseline materials as an input to EM-field based computational modelling and optimization, absorption by the studied metamaterials measured by their reflection loss (RL) increases significantly. The resulting metamaterials have the potential for lower cost and lighter weight while providing greater protection than traditional metal gaskets and foams.
The manufacturing process of all-solid-state batteries necessitates the use of polymer ***,these binders,being ionic insulators by nature,can adversely affect charge transport within composite cathodes,thereby impacti...
详细信息
The manufacturing process of all-solid-state batteries necessitates the use of polymer ***,these binders,being ionic insulators by nature,can adversely affect charge transport within composite cathodes,thereby impacting the rate performance of the *** this work,we aim to investigate the impact of fabrication methods,specifically the solvent-free dry process versus the slurry-cast wet process,on binder distribution and charge transport in composite cathodes of solid-state *** the dry process,the binder forms a fibrous network,while the wet process results in binder coverage on the surface of cathode active *** difference in microstructure leads to a notable 20-fold increase in ionic conductivity in the dry-processed ***,the cells processed via the dry method exhibit higher capacity retention of 89%and 83%at C/3 and C/2 rates,respectively,in comparison to 68%and 58%for the wet-processed cells at the same *** findings provide valuable insights into the influence of fabrication methods on binder distribution and charge transport,contributing to a better understanding of the binder’s role in manufacturing of all-solid-state batteries.
This paper presents a dendrite-like device that discriminates spatiotemporal patterns of pulses for parallel processing in 3D neuromorphic architectures. The device utilizes the ferroelectric layer in a segmented mult...
This paper presents a dendrite-like device that discriminates spatiotemporal patterns of pulses for parallel processing in 3D neuromorphic architectures. The device utilizes the ferroelectric layer in a segmented multi-gate FeFET design to detect a consecutive sequence of input pulses. Experimental results demonstrate successful emulation of highly selective sequence discrimination in dendrites of neurons in the cortex and highlight up to 100× signal-margin (output current differences). This nanodendrite design offers a neuromorphic solution to thermally scalable parallel processing in 3D systems.
Pet owners experience difficulty in understanding their pets' body language and its implications for animal welfare, given that animals cannot utilize human speech to communicate their emotions and health conditio...
Pet owners experience difficulty in understanding their pets' body language and its implications for animal welfare, given that animals cannot utilize human speech to communicate their emotions and health conditions. However, previous experiments for analyzing cat behavior have demonstrated that cats are precisely expressive. DeepCat, a deep-learning approach developed in this study, translates cats' body language signals, enabling owners to discern their feline companions' intended messages and emotional states. Our DeepCat model was trained on a dataset comprising 10,000 cat images, implementing automatic labeling to track key features, including the tail, eyes, and mouth. Presented as a Flutter application, DeepCat can function everywhere, allowing owners to easily monitor their cats and make informed decisions in situations that require caution. This paper discusses the potential benefits and limitations of DeepCat and provides suggestions for future research in this domain.
Bulk switching RRAM devices have emerged to address nonidealities of filamentary RRAM for AI at the edge. However, long retention and high endurance needed for continual on-chip learning have yet to be demonstrated. H...
详细信息
ISBN:
(数字)9798350365429
ISBN:
(纸本)9798350365436
Bulk switching RRAM devices have emerged to address nonidealities of filamentary RRAM for AI at the edge. However, long retention and high endurance needed for continual on-chip learning have yet to be demonstrated. Here, we report a filament-free and multilevel bulk RRAM (b-RRAM) technology with long retention and high endurance. The oxygen vacancy gradient across the $\text{TiO}_{\mathrm{x}}$ switching layer and $\text{Al}_{2}\mathrm{O}_{3}$ tunnel barrier are designed to suppress filament formation and achieve multilevel switching through modulation of oxygen vacancy distribution. AI-Ti-O oxygen barrier enables reliable switching with high uniformity, retention and endurance. We develop a compact model that captures both DC and pulse switching. $\mathrm{M}\Omega$ -level resistance and switching current nonlinearity allow highly accurate read/write operations and matrix vector multiplications (MVM) in selectorless b-RRAM crossbars. A neuro-inspired few-shot learning (FSL) algorithm based on dendritic computation and behavioral time plasticity (BTSP) is mapped onto b-RRAM crossbar arrays. Weights are continuously updated on-chip to learn shortest navigation to reward within a few trials in a 2D maze. We also present FSL of large-size images with hyperdimensional (HD) computing using our b-RRAM model. System-level simulations with 64-level b-RRAM arrays show an increase in accuracy by 5.9 % and 5.4 % for a 5-way 10-shot learning of CIFAR-100 and CUB-200 datasets as compared to the state-of-the-art FSL-HD.
暂无评论