Search

Representation learning is an important topic in machine learning, pattern recognition, and data mining research. Among many representation learning approaches, semi-nonnegative matrix factorization (SNMF) is a frequently-used one. However, a typical problem of SNMF is that usually there is no learning rate guidance during the optimization process, which often leads to a poor representation ability. To overcome this limitation, we propose a very general representation learning framework (DNSRF) that is based on a deep neural net. Essentially, the parameters of the deep net used to construct the DNSRF algorithms are obtained by matrix element update. In combination with different activation functions, DNSRF can be implemented in various ways. In our experiments, we tested nine instances of our DNSRF framework on six benchmark datasets. In comparison with other state-of-the-art methods, the results demonstrate the superior performance of our framework, which is thus shown to have a great representation ability. [ABSTRACT FROM AUTHOR]

Intelligent mobile robots have been gradually used in various fields, including logistics, healthcare, service, and maintenance. Path planning is a crucial aspect of intelligent mobile robot research, which aims to empower robots to create optimal trajectories within complex and dynamic environments autonomously. This study introduces an improved A* algorithm to address the challenges faced by the preliminary A* pathfinding algorithm, which include limited efficiency, inadequate robustness, and excessive node traversal. Firstly, the node storage structure is optimized using a minimum heap to decrease node traversal time. In addition, the heuristic function is improved by adding an adaptive weight function and a turn penalty function. The original 8-neighbor is expanded to a 16-neighbor within the search strategy, followed by the elimination of invalid search neighbor to refine it into a new 8-neighbor according to the principle of symmetry, thereby enhancing the directionality of the A* algorithm and improving search efficiency. Furthermore, a bidirectional search mechanism is implemented to further reduce search time. Finally, trajectory optimization is performed on the planned paths using path node elimination and cubic Bezier curves, which aligns the optimized paths more closely with the kinematic constraints of the robot derivable trajectories. In simulation experiments on grid maps of different sizes, it was demonstrated that the proposed improved A* algorithm outperforms the preliminary A* Algorithm in various metrics, such as search efficiency, node traversal count, path length, and inflection points. The improved algorithm provides substantial value for practical applications by efficiently planning optimal paths in complex environments and ensuring robot drivability. [ABSTRACT FROM AUTHOR]

Objective: To evaluate the safety study of percutaneous gastroscopic gastrostomy in patients after ventriculoperitoneal shunt. Methods: We conducted a retrospective analysis of neurosurgical patients who underwent VPS and PEG at our hospital between January 2012 and November 2023. Patients were divided into 2 groups: VPS group and VPS followed by PEG gruop. Patients received routine antibiotic prophylaxis before the procedure, continued for 48 hours. Follow-up included monitoring immediate complications, particularly wound infection, intracranial infection, neurologic status deterioration, and shunt dysfunction. Routine follow-up visits were conducted post-discharge. Results: In the VPS group (n = 778), the incidence of intracranial infection was 3.08%. Among patients with PEG after VPS, the time interval between procedures ranged from 13 to 685 days. The mean follow-up period was 22 (1-77) months, with no deaths or further complications. Conclusion: Performing PEG more than 13 days after VPS does not significantly increase the risk of intracranial infections or PEG-associated infections, making it a relatively safe procedure. [ABSTRACT FROM AUTHOR]

Liver fibrosis is one of the histopathological characters during Echinococcus multilocularis infection. The activation of hepatic stellate cells (HSCs) is a key event in the development of liver fibrosis. However, the molecular mechanism of HSC activation in the E. multilocularis infection-induced liver fibrosis remains largely unclear. Here, we reported that mmu-miR-342-3p was most dominantly expressed in HSCs and was upregulated in the HSCs in response to E. multilocularis infection. We further showed that mmu-miR-342-3p was able to bind to the 3' UTR of the Zbtb7a gene and regulated its expression. Moreover, mmu-miR-342-3p expression was negatively correlated with its target gene Zbtb7a in HSCs during E. multilocularis infection. Knockdown of mmu-miR-342-3p promoted the expression of Gfap in the activated HSCs in vitro. In the E. multilocularis-infected mice, knockdown of mmu-miR-342-3p suppressed the expression of α-Sma, Col1α1, and TGF-β but promoted the expression of Gfap. Therefore, mmu-miR-342-3p is a key regulator for activation of HSCs, and inhibiting mmu-miR-342-3p to promote Zbtb7a-mediated TGF-β signaling in activated HSCs could be a novel strategy to treat liver fibrosis induced by E. multilocularis. Author summary: Liver fibrosis is one of the histopathological features during Echinococcus multilocularis infection. The activation of hepatic stellate cells (HSCs) is a key event in the development of liver fibrosis. However, the molecular mechanism of HSC activation in the liver fibrosis induced by E. multilocularis infection remains largely unknown. In recent years, there is increasing evidence indicate that the abnormal expression of miRNAs is closely related to the occurrence and development of liver fibrosis. The expression profile of miRNA in hepatic fibrosis induced by E. multilocularis infection has been widely characterized, but the roles of miRNAs in liver fibrosis are largely unexplored. Herein, we identified a set of differentially expressed miRNAs in the activated HSCs induced by E. multilocularis, mmu-miR-342-3p of which was dominantly expressed in HSCs. This study signifies the important role of mmu-miR-342-3p in the HSC activation, which will help us to better understand the mechanism of liver fibrosis induced by E. multilocularis infection. [ABSTRACT FROM AUTHOR]