Dissolved Gas Analysis (DGA) in transformer oil is an important technical means for the operation and maintenance of transformers, and clustering algorithms are an important intelligent algorithm for oil chromatograph...
详细信息
Given the current challenging labor market, many organizations are grappling with significant issues related to employee turnover. this turnover can lead to considerable economic detriments for businesses, rendering t...
详细信息
Deepfake technology poses a significant threat to the integrity of multimedia content, raising urgent concerns for media authentication and trustworthiness. this paper addresses the challenges of detecting deep fake v...
详细信息
Zeroth-order (ZO) methods, which use the finite difference of two function evaluations (also called ZO gradient) to approximate first-order gradient, have attracted much attention recently in machine learning because ...
详细信息
Zeroth-order (ZO) methods, which use the finite difference of two function evaluations (also called ZO gradient) to approximate first-order gradient, have attracted much attention recently in machine learning because of their broad applications. the accuracy of the ZO gradient highly depends on how many finite differences are averaged, which are intrinsically determined by the number of perturbations randomly drawn from a distribution. Existing ZO methods try to learn a data-driven distribution for sampling the perturbations to improve the efficiency of ZO optimization (ZOO) algorithms. In this paper, we explore a new and parallel direction, i.e., learn an optimal sampling policy instead of using a random strategy to generate perturbations based on the techniques of reinforcement learning (RL), which makes it possible to approximate the gradient with only two function evaluations. Specifically, we first formulate the problem of learning a sampling policy as a Markov decision process. then, we propose our ZO-RL algorithm, i.e., using deep deterministic policy gradient, an actor-critic RL algorithm to learn a sampling policy that can guide the generation of perturbed vectors in getting ZO gradients as accurately as possible. Importantly, the existing ZOO algorithms for learning a distribution can be plugged in to improve the exploration of ZO-RL. Experimental results with different ZO estimators show that our ZO-RL algorithm can effectively reduce the query complexity of ZOO algorithms and converge faster than existing ZOO algorithms, especially in the later stage of the optimization process.
Aiming at the problem of the controller parameter perturbation and exponential stabilization of controlled plants in complex working environments, an elastic stabilization control method, which is based on polynomial ...
详细信息
Machine learning is the process of automating and analyzing patterns within data and using these identified patterns to make predictions. Supervised learning focuses on scenarios where a labeled dataset is used to tra...
详细信息
In this paper, we investigate a more efficient IoU loss based bounding box localization mechanism on top of end-to-end target detection frameworks to further improve the regression accuracy of object detection methods...
详细信息
ISBN:
(纸本)9798400710353
In this paper, we investigate a more efficient IoU loss based bounding box localization mechanism on top of end-to-end target detection frameworks to further improve the regression accuracy of object detection methods. Aiming at the limited spatial perception ability of deep convolution features, this paper proposes an optimization strategy by combining shallow spatial feature feed-forward mechanism (SSFF) with focal IoU loss function for bounding box regression tasks in target detection. this strategy firstly constructs a channel to transfer important location information in shallow spatial features to deep spatial features to reduce the loss of spatial details. Secondly, it constructs a focal IoU loss function based on the IoU loss, which dynamically adjusts loss weights according to the regression difficulty of different bounding boxes, to improve the positioning ability of regression networks on difficult bounding boxes. the experimental results show that the proposed methods can effectively improve the regression accuracy of target detection models.
How to plan the vehicle path efficiently and accurately is a key problem in route planning. To solve this problem, in this paper, an Algorithm combining Ant Colony Algorithm (ACA) and Proximal Policy optimization (Pro...
详细信息
In the current hybrid edge and cloud computing network, efficient heterogeneous resource scheduling is crucial for data security, compliance and business growth. However, Kubernetes' default scheduling strategy ha...
详细信息
the complexity of irrigation management necessitates the use of professional experience for effective decision-making. Expert systems (ES) are specifically engineered to mimic human expertise efficiently, allowing cog...
详细信息
ISBN:
(纸本)9783031686597;9783031686603
the complexity of irrigation management necessitates the use of professional experience for effective decision-making. Expert systems (ES) are specifically engineered to mimic human expertise efficiently, allowing cognitive mechanisms to be replicated at the irrigation level. the value of combining human knowledge with cutting-edge technologies such as fog computing, cloud computing, blockchain, IoT, and machine learning has recently come to be recognized more and more. the objective is to create an integrated platform that increases irrigation systems' resilience and adaptability while also increasing their accuracy, efficiency, and capacity for adaptive decision-making in light of agricultural environments' changing conditions. With a focus on ES, fuzzy logic, and machine learningalgorithms, this paper examines the research trends and applicability of intelligent approaches for optimal irrigation management scenarios. It also provides a thorough analysis of how these approaches contribute to the overall success of smart irrigation management.
暂无评论