Yang Zhao

Artificial intelligence (AI) technology has brought new opportunities and challenges to the judicial field, which dramatically improves judicial efficiency and may even change the judiciary's way. The concept of judicial justice in the information age has a natural affinity with artificial intelligence. As artificial intelligence continues to make breakthroughs in judicial data sorting and deep learning knowledge, judicial artificial intelligence has gradually become a reality. Artificial intelligence can conduct legal argumentation, interpret calculation results, human–computer collaboration, and judicial judgment. At the same time, the development of artificial intelligence technology also has potential risks, such as algorithm black boxes, algorithm discrimination, etc., to help judges achieve judicial justice, the value of judicial justice should be pursued to the limit of its judicial application. Therefore, artificial intelligence justice should be constrained by social relations, legal rationality, and code operation.

Background Community advisory boards have expanded beyond high-income countries and play an increasing role in low- and middle-income country research. Much research has examined CABs in HICs, but less is known about CABs in LMICs. The purposes of this scoping review are to examine the creation and implementation of CABs in LMICs, including identifying frequently reported challenges, and to discuss implications for research ethics. Methods We searched five databases for publications describing or evaluating CABs in LMICs. Two researchers independently reviewed articles for inclusion. Data related to the following aspects of CABs were extracted from included publications: time, country, financial support, research focus, responsibilities, and challenges. Thematic analyses were used to summarize textual data describing challenges. Results Our search yielded 2005 citations, 83 of which were deemed eligible for inclusion. Most studies were published between 2010 and 2017. Upper-middle-income countries were more likely to have studies describing CABs, with South Africa, China, and Thailand having the greatest numbers. The United States National Institutes of Health was the main source of financial support for CABs. Many CABs focused on HIV research. Thirty-four studies reported how CABs influenced the informed consent process for clinical trials or other aspects of research ethics. CAB responsibilities were related to clinical trials, including reviewing study protocols, educating local communities about research activities, and promoting the ethical conduct of research. Challenges faced by CABs included the following: incomplete ethical regulations and guidance; limited knowledge of science among members of communities and CABs; unstable and unbalanced power relationships between researchers and local communities; poor CAB management, including lack of formal participation structures and absence of CAB leadership; competing demands for time that limited participation in CAB activities; and language barriers between research staff and community members. Several challenges reflected shortcomings within the research team. Conclusions Our findings examine the formation and implementation of CABs in LMICs and identify several ethical challenges. These findings suggest the need for further ethics training among CAB members and researchers in LMICs.

As a part of user experience, user emotion has rarely been studied in mobile libraries. Specifically, with the proposed emotional model in combination with the Pleasure-Arousal-Dominance Emotion Model and the Five Factor Model, we evaluate user emotions on the mobile library’s three IS features. An experience procedure with three tasks has been designed to collect data. 50 participants were enrolled, and they were asked to fill in questionnaires right after the experience. The correlations among the PAD emotions were examined. Specifically, users have a low perception of pleasure, high perception of arousal, and low perception of dominance. However, these three emotional states were not always significantly correlated with each other. This study extends mobile library research by focusing on users’ emotional experience. Specifically, the detailed PAD emotions have been examined. This study provides a new approach for application developers and managers to evaluate the user experience of an application.

Based on the existing financial system risk models, a set of time-lag financial system risk models is established considering the influence brought by time-lag factors on the financial risk system, and the dynamical behavior of this system is analyzed by using chaos theory. Through Matlab simulation, the bifurcation diagram and phase diagram of time-lag risk intensity and control intensity are plotted. The analysis shows that this kind of time-lag financial system risk model has complex dynamic behavior, different motion states will appear when different parameter values are selected, and the time-lag risk intensity parameter also has a very strong influence on the system motion. To ensure the operation of the financial system in a stable state, measures with certain delay effects must be taken to control the risk and to choose the appropriate time-lag control intensity, and too much or too little time-lag control intensity is not conducive to the benign operation of the system.

Person reidentification aims to recognize a specific pedestrian from uncrossed surveillance camera views. Most re-id methods perform the retrieval task by comparing the similarity of pedestrian features extracted from deep learning models. Therefore, learning a discriminative feature is critical for person reidentification. Many works supervise the model learning with one or more loss functions to obtain the discriminability of features. Softmax loss is one of the widely used loss functions in re-id. However, traditional softmax loss inherently focuses on the feature separability and fails to consider the compactness of within-class features. To further improve the accuracy of re-id, many efforts are conducted to shrink within-class discrepancy as well as between-class similarity. In this paper, we propose a circle-based ratio loss for person re-identification. Concretely, we normalize the learned features and classification weights to map these vectors in the hypersphere. Then we take the ratio of the maximal intraclass distance and the minimal interclass distance as an objective loss, so the between-class separability and within-class compactness can be optimized simultaneously during the training stage. Finally, with the joint training of an improved softmax loss and the ratio loss, the deep model could mine discriminative pedestrian information and learn robust features for the re-id task. Comprehensive experiments on three re-id benchmark datasets are carried out to illustrate the effectiveness of the proposed method. Specially, 83.12% mAP on Market-1501, 71.66% mAP on DukeMTMC-reID, and 66.26%/63.24% mAP on CUHK03 labeled/detected are achieved, respectively.

The traction forces exerted by an adherent cell on a substrate have been studied only in the two-dimensions tangential to substrate surface. We developed a novel technique to measure the three-dimensional traction forces exerted by live bovine aortic endothelial cells on polyacrylamide deformable substrate. On 3D images acquired by confocal microscopy, displacements were determined with image-processing programs, and traction forces in tangential and normal directions were computed by finite element method. BAECs generated traction force in normal direction with an order of magnitude comparable to Txy. Tz is upward at the cell edge and downward under the nucleus, changing continuously with a sign reversal between cell edge and nucleus edge. The method was evaluated regarding accuracy and precision of displacement measurements, effects of FE mesh size, displacement noises, and simple bootstrapping. These results provide new insights into cell-matrix interactions in terms of spatial and temporal variations in traction forces in 3D. This technique can be applied to study live cells to assess their biomechanical dynamics in conjunction with biochemical and functional activities, for investigating cellular functions in health and disease.