A novel intrusion detection model for the CAN bus packet of in-vehicle network based on attention mechanism and autoencoder

The attacks on in-vehicle Controller Area Network(CAN)bus messages severely disrupt normal communication between vehicles.Therefore,researches on intrusion detection models for CAN have positive business value for vehicle security,and the intrusion detection technology for CAN bus messages can effectively protect the invehicle network from unlawful attacks.Previous machine learning-based models are unable to effectively identify intrusive abnormal messages due to their inherent shortcomings.Hence,to address the shortcomings of the previous machine learning-based intrusion detection technique,we propose a novel method using Attention Mechanism and AutoEncoder for Intrusion Detection(AMAEID).The AMAEID model first converts the raw hexadecimal message data into binary format to obtain better input.Then the AMAEID model encodes and decodes the binary message data using a multi-layer denoising autoencoder model to obtain a hidden feature representation that can represent the potential features behind the message data at a deeper level.Finally,the AMAEID model uses the attention mechanism and the fully connected layer network to infer whether the message is an abnormal message or not.The experimental results with three evaluation metrics on a real in-vehicle CAN bus message dataset outperform some traditional machine learning algorithms,demonstrating the effectiveness of the AMAEID model.

An Investigation into Disease Progression and the Correlation between Diabetes Mellitus and Hepatocellular Carcinoma

Diabetes mellitus has been associated with anincreased risk of hepatocellular carcinoma in recent studies ofpatients. At the same time, advanced hepatocellular carcinomaitself can cause glucose intolerance and can aggravate diabetes.Diabetes mellitus inducing hepatocellular carcinoma may resultin changes in the following aspects: dysfunction of organism,endocrine hormone balance and interactions, endothelins andso on. One way, diabetes mellitus may induce hepatocellularcarcinoma through the effects of chemotherapeutics and otheradjuvant drugs. This review outlines the relationship betweendiabetes mellitus and the risk of hepatocellular carcinoma as wellas treatments for hepatocellular carcinoma, which may be helpfulfor clinicians.

Generation of 84-fs pulses from a mode-locked Tm：CNNGG disordered garnet crystal laser

A mode-locked laser based on a Tm:CNNGG disordered crystal as an active medium and a single-walled carbon nanotube saturable absorber is demonstrated, operating at a central wavelength of 2018 nm. Transformlimited 84 fs pulses are generated with an average output power of 22 mW at a repetition rate of ~90 MHz.A maximum output power of 98 mW is obtained at a slightly longer pulse duration of 114 fs.

Exploring the genetic characteristics of 93-11 and Nipponbare recombination inbred lines based on the Golden Gate SNP assay

Understanding genetic characteristics in rice populations will facilitate exploring evolutionary mechanisms and gene cloning. Numerous molecular markers have been utilized in linkage map construction and quantitative trait locus （QTL） mappings. However, segregation-distorted markers were rarely considered, which prevented understanding genetic characteristics in many populations. In this study, we designed a 384-marker GoldenGate SNP array to genotype 283 recombination inbred lines （RILs） derived from 93-11 and Nipponbare Oryza sativa crosses. Using 294 markers that were highly polymorpbic between parents, a linkage map with a total genetic distance of 1,583.2 cM was constructed, including 231 segregation-distorted mark- ers. This linkage map was consistent with maps generated by other methods in previous studies. In total, 85 significant quanti- tative trait loci （QTLs） with phenotypic variation explained （PVE） values〉5% were identified. Among them, 34 QTLs were overlapped with reported genes/QTLs relevant to corresponding traits, and 17 QTLs were overlapped with reported sterili- ty-related genes/QTLs. Our study provides evidence that segregation-distorted markers can be used in linkage map construc- tion and QTL mapping. Moreover, genetic information resulting from this study will help us to understand recombination events and segregation distortion. Furthermore, this study will facilitate gene cloning and understanding mechanism of in- ter-subspecies hybrid sterility and correlations with important agronomic traits in rice.