Search

Therapeutic tumor vaccines have become an important breakthrough in the treatment of various solid tumors including lung cancer. Dendritic cells (DCs)-based tumor vaccines targeting tumor-associated antigens (TAAs) play a key role in immunotherapy and immunoprevention. However, the weak immunogenicity of TAAs and low immune response rates are a major challenge faced in the application of therapeutic tumor vaccines. Here, we tested whether targeting an attractive target Mesothelin (MSLN) and PD-L1 immune checkpoint molecule to DCs in vivo would elicit therapeutic antitumor cytotoxic T lymphocyte (CTL) response. We generated specific MSLN fragment combined with PD-L1 and GM-CSF peptide immunogen (MSLN-PDL1-GMCSF) based on the novel anti-PD-L1 vaccination strategy we recently developed for the cancer treatment and prevention. We found that DCs loaded with MSLN-PDL1-GMCSF vaccine elicited much stronger endogenous anti-PD-L1 antibody and T cell responses in immunized mice and that antigen specific CTLs had cytolytic activities against tumor cells expressing both MSLN and PD-L1. We demonstrated that vaccination with MSLN-PDL1-GMCSF potently inhibited the tumor growth of MSLN+ and PD-L1+ lung cancer cells, exhibiting a significant therapeutic anti-tumor potential. Furthermore, PD-1 blockade further improved the synergistic antitumor therapeutic efficacy of MSLN-PDL1-GMCSF vaccine in immunized mice. In summary, our data demonstrated for the first time that this PD-L1-containing MSLN therapeutic vaccine can induce persistent anti-PD-L1 antibody and CTL responses, providing an effective immunotherapeutic strategy for lung cancer immunotherapy by combining MSLN-PDL1-GMCSF vaccine and PD-1 blockade.

With the development of science and technology, the manufacturing technology of computer numerical control (CNC) has been one of the irreplaceable important technologies in the manufacturing industry. However, due to the lack of sufficient data, the existing evaluation methods cannot accurately evaluate the operational state of CNC machine tools, and an effective maintenance scheduling strategy cannot be proposed. Aiming at the lack of monitoring data and the unclear understanding of the degradation mechanism of CNC machine tools, an operational state evaluation and maintenance decision-making method of multi-state CNC machine tools based on partially observable Markov decision process (POMDP) was proposed. On the one hand, considering the imperfect operational information of CNC machine tools and the influence of different maintenance activities on the degradation process of NC machine tools, a dynamic evaluation model for the operational state was constructed. On the other hand, a maintenance decision model for CNC machine tools was established with the goal of minimizing the total cost of losses (including maintenance costs and potential failures costs). The proposed method is applied to CNC machine tools of cylinder to verify the effectiveness and advancement of the method.

This paper proposed a new fault diagnosis method based on the fault phenomenon description text. This paper used feature extraction algorithm in the field of natural language process to transform unstructured text data into structured data. In this paper, labeled text data is put into a support vector machine to train a prediction model. The model uses fault text description data to diagnose fault.

As modern industrial systems become more complex, intelligent, and large-scale, the reliability and safety of the system become more and more urgent. The basis for effectively ensuring the stable operation of complex systems lies in accurately describing the degradation law of the system’s health state, and then evaluating the health state of the system. However, existing research on system health assessment ignores the dependencies between components. Aiming at the difficulty of accurately modeling the health state of complex systems, a health state assessment model based on gray clustering and entropy weights was proposed in this paper. Firstly, based on historical data and expert experience, the key parameters that characterize the health of the system were determined. Secondly, the entropy weight method is used to calculate the weights of various parameters that characterize the health state of the system. On the basis, the health state evaluation result of complex systems is obtained based on the gray clustering method. Finally, taking the naval gun engine as the research object, the correctness and effectiveness of the proposed model were verified, and the results show that this method can effectively evaluate the health state of the system.

Network embedding, mapping nodes in a network to a low-dimensional space, achieves powerful performance. An increasing number of works focus on static network embedding, however, seldom attention has been paid to temporal network embedding, especially without considering the effect of mesoscopic dynamics when the network evolves. In light of this, we concentrate on a particular motif --- triad --- and its temporal dynamics, to study the temporal network embedding. Specifically, we propose MTNE, a novel embedding model for temporal networks. MTNE not only integrates the Hawkes process to stimulate the triad evolution process that preserves motif-aware high-order proximities, but also combines attention mechanism to distinguish the importance of different types of triads better. Experiments on various real-world temporal networks demonstrate that, compared with several state-of-the-art methods, our model achieves the best performance in both static and dynamic tasks, including node classification, link prediction, and link recommendation.

A nitrogen plasma treatment protocol has been applied to carbon nanotubes as a way to modulate their magnetic properties by the microwave plasma chemical vapor deposition method. The nitrogen treatment is a simple and effective strategy to fundamentally affect the magnetic performance of CNTs, which is related to the microwave power and the time of the treatment. The best treatment time is 30 min with the microwave power of 120 W corresponding to the N content (defined N/C atomic ratio in at.%) which is 2.15 at.% and the biggest value of Ms is 0.5227 emu/g. Using Raman, XPS and EDS mapping, it is proved that N exists in CNTs and the pyridine N and amine N bonding configurations contribute to ferromagnetic origin. Our study of N2 plasma treatment doping into CNTs to enhance ferromagnetic properties provides an effective technique for the N doping and massive synthesis of ferromagnetic graphene.

This paper proposes a combination detection and fusion method for targets on sea surface, which is based on radar and infrared channels, this method is also suitable for sea surface target combination detection and fusion of radar and visible light camera at the same time. Radar detection has the advantages of full-time, all-weather, and large detection range, but the ability to acquire target details is poor. In comparison with radar, infrared detection has higher angular resolution and can acquire the target's thermal radiation characteristics, which allows for a wealth of detailed information. However, the infrared field of view is small. Based on the advantages and disadvantages of these two detection methods, the radar is used for target detection. Then, according to the azimuth angle and the distance of the target in the radar image, the infrared camera is guided to observe the target. The combined detection method of radar-guide-infrared can combine the advantages of radar’s long detection range, large coverage area with the high resolution and rich detail information of infrared, which covers the details missing of radar detection and the small view field of infrared. After the radar directs the infrared to the target, we will get the infrared image of target. If there is more than one target, it is necessary to merge the data of the radar channel and infrared channel, so that the targets detected by the radar correspond to the targets of the infrared detection. However, due to the different error sources of radar and infrared detection, the error characteristics of the two channels are greatly different, which makes the target fusion of radar and infrared channel difficult. In order to deal with this problem, first of all, the Extend Kalman filter models for radar channel and infrared channel are designed respectively, which are on the base of the imaging characteristics of radar and infrared. After the filtering of the target position information obtained from radar and infrared channels, the random error of the target and the error of target solution will be reduced. What’s more, the velocity information of the targets will be solved at the same time. After that, the error distribution mathematical models of the radar and the infrared channels will be established. Combine the error distribution models with the filtering results and the relationship between the radar coordinate system and the infrared coordinate system, the probability distribution of the targets in radar image and infrared image will be obtained. Finally, according to the probability of the target distribution, numerical method is used to compute the matching matrix of the targets in radar and infrared images. After dealing with the matching matrix, the target fusion of radar channel and infrared channel is accomplished. After that, according to the target matching result, the radar and infrared data are fused to obtain higher-precision target position information.