Sharpness-Aware Minimization (SAM), which performs gradient descent on adversarially perturbed weights, can improve generalization by identifying flatter minima. However, recent studies have shown that SAM may suffer ...
Sharpness-Aware Minimization (SAM), which performs gradient descent on adversarially perturbed weights, can improve generalization by identifying flatter minima. However, recent studies have shown that SAM may suffer from convergence instability and oscillate around saddle points, resulting in slow convergence and inferior performance. To address this problem, we propose the use of a lookahead mechanism to gather more information about the landscape by looking further ahead, and thus find a better trajectory to converge. By examining the nature of SAM, we simplify the extrapolation procedure, resulting in a more efficient algorithm. Theoretical results show that the proposed method converges to a stationary point and is less prone to saddle points. Experiments on standard benchmark datasets also verify that the proposed method outperforms the SOTAs, and converge more effectively to flat minima.
The development of intelligent street light systems has ushered in a new era of efficiency and sustainability in urban infrastructure. The proposed work studies the integration of modern sensors and Internet of Things...
详细信息
ISBN:
(数字)9798350384369
ISBN:
(纸本)9798350384376
The development of intelligent street light systems has ushered in a new era of efficiency and sustainability in urban infrastructure. The proposed work studies the integration of modern sensors and Internet of Things (IoT) technologies to increase the functionality of street lights. The major components include Light Dependent Resistors (LDR) for automatic light management, an MQ135 sensor for air quality monitoring, and a fog detection system employing a laser module and LDR. These sensors are integrated with the ESP8266, providing cloud integration for real-time data visualization and monitoring. Additionally, solar panels are employed to enhance energy efficiency. The proposed methodology digs into the precise implementation, data analysis, and the potential of this system to alter urban lighting infrastructure
Currently, Digital Holographic Microscopy(DHM) is used for researching disease diagnosis or microbes. It cannot obtain the correct three-dimensional (3D) profile for a little noise because biological cells are microsc...
详细信息
This paper presents the design of a system onboard spider robot for rescue services. The spider has the six-leg walking pattern that contains the total number of 18 servo motors. In order to reduce the payload, the ab...
详细信息
Sesame (Sesamum indicum L.) is an important commercial and food crop, and its yields is limited by many insects, pests, diseases, and weeds. Autonomous aerial agrochemicals spray application on sesame fields using a d...
详细信息
ISBN:
(纸本)9781665498197
Sesame (Sesamum indicum L.) is an important commercial and food crop, and its yields is limited by many insects, pests, diseases, and weeds. Autonomous aerial agrochemicals spray application on sesame fields using a drone aims to save crop from these yield limiting factors and in addition agrochemicals application quantity and site could be controlled, and human health is expected to be protected. For accurate and selective spray application, autonomous systems would need some parameters to distinguish between crop, weed and background. In this research an aerial sesame field dataset has been collected with the focus to classify patch areas of sesame and weeds present in the field. Dataset was captured using Agrocam. We have developed a patch image-based classification approach along with a novel SesameWeedNet convolutional neural network (CNN) inspired by the layer's configuration of VGG networks and depth-wise convolutions of the MobileNet. The small model contains 6 convolutional layers, and it runs faster and accurately on small patch images. Our approach breaks 1920×1080-pixel images into smaller patch images of size 45×45 pixels. After that, these small patch images are fed to a relatively small CNN for training, validation, and finally for classification. Patch based model ensemble and dataset grouping are two major parts in our methodology. Our system recommends the dataset grouping according to vegetation present in the images to enhance classification results. We have achieved accuracy up to 96.7% with our proposed method. We have tested our system under sunlight variation, in wet and dry soil conditions and at different growth stages. To the best of our knowledge, no attempt has been made to classify and treat crop and weeds in sesame fields at the post-emergence stage previously. In this research we have made the contribution of aerial sesame-weed dataset and a complete deep learning-based approach to classify weeds in sesame fields under variable lighti
This study explores the effectiveness of Convolutional Neural Networks (CNNs) in automatically classifying skin cancer for e-health applications. The trained model showcases impressive performance by leveraging the HA...
详细信息
In this paper, we propose the depth mapping and occlusion removal method using integral imaging (InIm). InIm is a method to obtain 3D information by capturing and reconstructing an object from multiple viewpoints. InI...
详细信息
Multipliers can be used to guarantee both the Lyapunov stability and input-output stability of Lurye systems with time-invariant memoryless slope-restricted nonlinearities. If a dynamic multiplier is used there is no ...
In this paper, we propose a method for visualizing object information under low light conditions using photon-counting integral imaging and depth images as prior information. To visualize 3D objects under low-light co...
详细信息
Artificial Intelligence (AI) and marketing have transformed consumer behavior and shopping experiences, especially through Recommender Systems (RSs) in e-commerce. RSs use algorithms to provide personalized recommenda...
详细信息
暂无评论