Perceptual quality assessment of the videos acquired in the wilds is of vital importance for quality assurance of video services. The inaccessibility of reference videos with pristine quality and the complexity of authentic distortions pose great challenges for this kind of blind video quality assessment (BVQA) task. Although model-based transfer learning is an effective and efficient paradigm for the BVQA task, it remains to be a challenge to explore what and how to bridge the domain shifts for better video representation. In this work, we propose to transfer knowledge from image quality assessment (IQA) databases with authentic distortions and large-scale action recognition with rich motion patterns. We rely on both groups of data to learn the feature extractor and use a mixed list-wise ranking loss function to train the entire model on the target VQA databases. Extensive experiments on six benchmarking databases demonstrate that our method performs very competitively under both individual database and mixed databases training settings. We also verify the rationality of each component of the proposed method and explore a simple ensemble trick for further improvement.

Reasonable repair crew dispatch can accelerate the service restoration in distribution systems after some components are damaged. It is worthy noting that as traditional distribution systems are involving into cyber-physical distribution systems composed of physical distribution networks and cyber networks, damaged components in these two networks should be coordinately repaired to achieve maximal restoration performance. However, repair crews in physical distribution networks are scheduled without considering the impact of the corresponding cyber networks in existing repair crew dispatch strategies. In this paper, a coordinated model is proposed to co-optimize the repair crew schedule of the two networks. Besides the cyber-physical interdependence between the two networks, backup power systems, topology reconfiguration and intentional distribution generator (DG) island are also modelled to improve the repair performance. The proposed model is formulated as a mixed-integer linear programming model, and its effectiveness is verified on the modified IEEE 33-node and 123-node test systems. Numerical results highlight the necessity of co-optimizing repair crew dispatch in both physical distribution networks and cyber networks.

This article provides a framework for modeling and analysis of spatial cyber-physical power systems (CPPSs). Previous studies on structural vulnerability analysis of CPPS mainly focused on network topology without any spatial information. In this article, we first propose a hierarchical information system evolution model for spatial power grids. Then, a spatial CPPS model is established by combining power and information systems that accounts for overlapped power and information lines. Finally, the structural vulnerability under global and localized failures is investigated based on direct current (dc) power flow. Under global failures, different types of components have distinguishing effects, and the power-node failure is the most serious case. For localized failures, key zones can be identified by the cascading failure scale, and similarly, different failure modes affect different areas. Furthermore, considering power system scenarios with only defenders and those with both attackers and defenders, we investigate two resource allocation situations. For defenders, increased resources and appropriately dispersed allocation can work effectively. However, in the attack–defense scenario, the perfect subgame equilibrium suggests that both attackers and defenders must concentrate all their resources on the most vulnerable zone. These findings may establish a foundation for future efforts to model and protect spatial CPPSs.