Class imbalance is an important factor that affects the performance of deep learning models used for remote sensing scene classification. In this paper, we propose a random finetuning meta metric learning model (RF-MML) to address this problem. Derived from episodic training in meta metric learning, a novel strategy is proposed to train the model, which consists of two phases, i.e., random episodic training and all classes fine-tuning. By introducing randomness into the episodic training and integrating it with fine-tuning for all classes, the few-shot meta-learning paradigm can be successfully applied to class imbalanced data to improve the classification performance. Experiments are conducted to demonstrate the effectiveness of the proposed model on class imbalanced datasets, and the results show the superiority of our model, as compared with other state-of-the-art methods.
Inefficient data transmission has been a development bottleneck of Vehicular Ad-hoc NETworks (VANETs), especially in urban areas. It has been proved that many complex IP-based solutions are difficult to be applied in the highly dynamic and link-interrupted vehicular environment. In recent years, Named Data Networking (NDN) has become the most popular realization of Information-Centric Networking (ICN) for future networks. Its characteristics of multi-source, multi-path and in-network caching are helpful for improving the data transmission in VANETs. However, the bottom layer of vehicles cannot provide multiple interfaces to different domains like the routers in wired networks. Thus interface-based forwarding degenerates into directionless broadcasting with low performance and high overhead. Against this problem, we propose COMPASS, a novel named data transmission protocol for VANETs. Firstly, a dynamic directional interface model is built as the cornerstone of our COMPASS. Secondly, the forwarding strategies are improved for rapid interest dissemination and named data retrieving. Besides, an interface remapping method and the update strategies of Forwarding Information Base (FIB) and Pending Interest Table (PIT) are designed to enhance the robustness in high-mobility environment. Finally, the performance of COMPASS is verified on ndnSIM. Compared with the other three state-of-the-art protocols, COMPASS obtains the highest interest satisfaction ratio and the shortest transmission delay in urban traffic scenarios with restricted communication and storage overhead. INDEX TERMS Vehicular named data networking (VNDN), named data transmission, forwarding strategy, dynamic directional interface.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.