The edge computing in knowledge-defined network (KDN) is a kind of distributed computing architecture, and the format of edge resources stored in different edge computing nodes are different, which yields the data heterogeneity problem and hampers the interaction between edge nodes. Ontology is considered as the solution of data heterogeneity on Semantic Web, and matching ontologies is a high-efficiency method of addressing the data heterogeneity problem. Ontology meta-matching investigates how to determine the optimal weights to aggregate multiple similarity measures to achieve high-quality ontology alignment, which is a challenge about nonlinear mathematical problem in ontology matching domain. To face this challenge, unsupervised learning method such as generative adversarial network (GAN) becomes an effective methodology. GAN consists of two models of different targets that are opposed to each other in training to produce the final best result. To improve the GAN's efficiency, this work further proposes a GAN with simulated annealing algorithm (SA-GAN), where the stagnation counter is introduced to accelerate GAN's the convergence speed. The experiment uses the famous benchmark in the ontology domain, and the comparisons with the advanced ontology matching systems shows that SA-GAN is able to find highquality alignments to help build bridges between edge nodes on edge computing.

Cisplatin is one of the most active chemotherapy drugs to treat solid tumors. However, it also causes various side effects, especially nephrotoxicity, in which oxidative stress plays critical roles. Our previous studies found that cisplatin selectively inhibited selenoenzyme thioredoxin reductase1 (TrxR1) in the kidney at an early stage and, subsequently, induced the activation of Nrf2. However, the effects of selenium on cisplatin-induced nephrotoxicity are still unclear. In this study, we established mice models with different selenium intake levels to explore the effects of selenoenzyme activity changes on cisplatin-induced nephrotoxicity. Results showed that feeding with a selenium-deficient diet sensitize the mice to cisplatin-induced damage, whereas selenium supplementation increased the activities of selenoenzymes TrxR and glutathione peroxidase (GPx), changed the renal cellular redox environment to a reduced state, and exhibited protective effects. These results demonstrated the correlation of selenoenzymes with cisplatin-induced side effects and provided a basis for the potential approach to alleviate cisplatin-induced renal injury.

Object detection is a fundamental task in computer vision and image processing. Current deep learning based object detectors have been highly successful with abundant labeled data. But in real life, it is not guaranteed that each object category has enough labeled samples for training. These large object detectors are easy to overfit when the training data is limited. Therefore, it is necessary to introduce few-shot learning and zeroshot learning into object detection, which can be named low-shot object detection together. Low-Shot Object Detection (LSOD) aims to detect objects from a few or even zero labeled data, which can be categorized into few-shot object detection (FSOD) and zero-shot object detection (ZSD), respectively. This paper conducts a comprehensive survey for deep learning based FSOD and ZSD. First, this survey classifies methods for FSOD and ZSD into different categories and discusses the pros and cons of them. Second, this survey reviews dataset settings and evaluation metrics for FSOD and ZSD, then analyzes the performance of different methods on these benchmarks. Finally, this survey discusses future challenges and promising directions for FSOD and ZSD.

No abstract

No abstract