Unmanned Aerial Vehicles (UAVs) are widely used in various applications, from inspection and surveillance to transportation and delivery. Navigating UAVs in complex 3D environments is a challenging task that requires ...
详细信息
Training agents that are robust to environmental changes remains a significant challenge in deep reinforcement learning (RL). Unsupervised environment design (UED) has recently emerged to address this issue by generat...
ISBN:
(纸本)9798331314385
Training agents that are robust to environmental changes remains a significant challenge in deep reinforcement learning (RL). Unsupervised environment design (UED) has recently emerged to address this issue by generating a set of training environments tailored to the agent's capabilities. While prior works demonstrate that UED has the potential to learn a robust policy, their performance is constrained by the capabilities of the environment generation. To this end, we propose a novel UED algorithm, adversarial environment design via regret-guided diffusion models (ADD). The proposed method guides the diffusion-based environment generator with the regret of the agent to produce environments that the agent finds challenging but conducive to further improvement. By exploiting the representation power of diffusion models, ADD can directly generate adversarial environments while maintaining the diversity of training environments, enabling the agent to effectively learn a robust policy. Our experimental results demonstrate that the proposed method successfully generates an instructive curriculum of environments, outperforming UED baselines in zero-shot generalization across novel, out-of-distribution environments. Project page: https://***/projects/ADD
Partial discharge (PD) is a widespread phenomenon instigated in power transformer (PT) insulation systems. PDs are triggered by voids that vary in size and position within the PT insulation. The electrical characteris...
详细信息
The amount of data processed in the cloud, the development of Internet-of-Things (IoT) applications, and growing data privacy concerns force the transition from cloud-based to edge-based processing. Limited energy and...
详细信息
This article examines the incorporation of the Shopping Assistance Automatic Suggestion (SAAS) model into Virtual Reality (VR) environments in order to improve the online shopping experience. The SAAS model employs so...
详细信息
One of the most important tasks in autonomous driving and autonomous vehicle navigation is detecting a path or trajectory that the vehicle should follow. Over the past few years, some learning-based works have stood o...
详细信息
ISBN:
(数字)9798350358513
ISBN:
(纸本)9798350358520
One of the most important tasks in autonomous driving and autonomous vehicle navigation is detecting a path or trajectory that the vehicle should follow. Over the past few years, some learning-based works have stood out more than traditional computer vision techniques in detecting such lanes. In this paper we present an approach to solve the lane line detection problem in the context of visual path following by using a residual factorized convolutional neural network. Experimental results show a promising model that can detect lane lines even under severe lighting conditions and in the presence of occlusions and shadows. The path detection system was tested along with a visual path following formulation based on Nonlinear Model Predictive Control. Still, it can be used for any controller in the context of visual navigation for autonomous vehicles. Nonetheless, the proposed model architecture strikes a remarkable balance between accuracy and efficiency, making the system suitable for real-time applications.
Several recent studies have elucidated why knowledge distillation (KD) improves model performance. However, few have researched the other advantages of KD in addition to its improving model performance. In this study,...
详细信息
This paper presents a trajectory planning and obstacle avoidance system for Unmanned Surface Vehicles (USV) in complex and dynamic navigation environments. The developed system employs a modified Artificial Potential ...
详细信息
ISBN:
(数字)9798350352344
ISBN:
(纸本)9798350352351
This paper presents a trajectory planning and obstacle avoidance system for Unmanned Surface Vehicles (USV) in complex and dynamic navigation environments. The developed system employs a modified Artificial Potential Field (APF) algorithm and the International Regulations for Preventing Collisions at Sea (COLREGS) to ensure safe and efficient navigation. Modifications were implemented to the original algorithm to deal with obstacles approaching the autonomous vehicle, including adding vectors to determine the direction of deviation based on the cross-product. The proposed system was validated in the Gazebo simulation environment within a dynamic scenario featuring static and moving obstacles.
Recently, the research on daily health monitoring using a wearable sensor has been continually evolving. In the future, when this system is actually implemented, a vast amount of data transmission will be conducted fr...
详细信息
作者：
Janet Van NiekerkHåvard RueStatistics Program
Computer Electrical and Mathematical Sciences and Engineering Division King Abdullah University of Science and Technology (KAUST) Thuwal Kingdom of Saudi Arabia
Approximate inference methods like the Laplace method, Laplace approximations and variational methods, amongst others, are popular methods when exact inference is not feasible due to the complexity of the model or the...
详细信息
Approximate inference methods like the Laplace method, Laplace approximations and variational methods, amongst others, are popular methods when exact inference is not feasible due to the complexity of the model or the abundance of data. In this paper we propose a hybrid approximate method called Low-Rank Variational Bayes correction (VBC), that uses the Laplace method and subsequently a Variational Bayes correction in a lower dimension, to the joint posterior mean. The cost is essentially that of the Laplace method which ensures scalability of the method, in both model complexity and data size. Models with fixed and unknown hyperparameters are considered, for simulated and real examples, for small and large data sets.
暂无评论