Graph structured data often possess dynamic characters in nature, such as the addition of links and nodes, in many real-world applications. Recent years have witnessed the increasing attentions paid to dynamic graph neural networks for modelling graph data. However, almost all existing approaches operate under the assumption that, upon the establishment of a new link, the embeddings of the neighboring nodes should undergo updates to learn temporal dynamics. Nevertheless, these approaches face the following limitation: If the node introduced by a new connection contains noisy information, propagating its knowledge to other nodes becomes unreliable and may even lead to the collapse of the model. In this paper, we propose Ada-DyGNN : a robust knowledge Ada ptation framework via reinforcement learning for Dy namic G raph N eural N etworks. In contrast to previous approaches, which update the embeddings of the neighbor nodes immediately after adding a new link, Ada-DyGNN adaptively determines which nodes should be updated. Considering that the decision to update the embedding of one neighbor node can significantly impact other neighbor nodes, we conceptualize the node update selection as a sequence decision problem and employ reinforcement learning to address it effectively. By this means, we can adaptively propagate knowledge to other nodes for learning robust node embedding representations. To the best of our knowledge, our approach constitutes the first attempt to explore robust knowledge adaptation via reinforcement learning specifically tailored for dynamic graph neural networks. Extensive experiments on three benchmark datasets demonstrate that Ada-DyGNN achieves the state-of-the-art performance. In addition, we conduct experiments by introducing different degrees of noise into the dataset, quantitatively and qualitatively illustrating the robustness of Ada-DyGNN.

Pollution flashover of insulators has been viewed as a common issue existing in the area of power system for a long term. However, traditional detection methods such as equivalent salt density and leakage current can hardly satisfy the requirements of the progressing smart grid. Herein, a novel highly-efficient method is proposed for classifying the pollution degree detection of insulators based on hyperspectral technology. The insulating sheets and actual insulators were set as the research object and their hyperspectral images are obtained. Moreover, the data for describing the characteristic wavelengths and the G color components of the samples were extracted from the images as the characterization of the equivalent salt deposit density (ESDD) and non-soluble deposit density (NSDD) respectively. The ESDD and NSDD detection model is established based on the multiple linear regression (MLR) algorithm and the random forest (RF) algorithm. Ultimately, the proposed model reaches detection accuracy of above 84% on the insulating sheets and 75% on the actual insulators, which can be viewed as acceptable. The hyperspectral imaging technique has considerable potential for the non-contact detection of ESDD and NSDD.

Off-policy evaluation is an important topic in reinforcement learning, which estimates the expected cumulative reward of a target policy using logged trajectory data generated from a different beha... In a sequential decision-making problem, off-policy evaluation estimates the expected cumulative reward of a target policy using logged trajectory data generated from a different behavior policy, w...