Search

Small and medium-sized enterprises currently suffer from a lack of branding. Therefore, to further promote their active branding, this study proposes a user preference mining algorithm based on data mining and social behavior. Employing this algorithm to study the degree of users’ brand preference can provide data support for enterprises’ brand building. The experimental results showed that the proposed algorithm outperforms previous algorithms in terms of performance, convergence, and accuracy. The area under the curve reached 0.953, indicating highly authentic output results with extremely high realism. In actual simulation experiments, its prediction results for the user’s brand preference index are accurate, with an error of only 0.11, and the algorithm has extremely high ratings among industry insiders. In conclusion, the user-preference mining algorithm based on data mining and social behaviors suggested in this study plays a better role in promoting an enterprise’s brand building. It can help the enterprise know the level of consumer preference for its brand; accordingly, it can determine the shortcomings in, provide effective and accurate data support for, and thereby promote its brand building.

Artificial intelligence (AI) holds significant promise in transforming medical imaging, enhancing diagnostics, and refining treatment strategies. However, the reliance on extensive multicenter datasets for training AI models poses challenges due to privacy concerns. Federated learning provides a solution by facilitating collaborative model training across multiple centers without sharing raw data. This study introduces a federated attention-consistent learning (FACL) framework to address challenges associated with large-scale pathological images and data heterogeneity. FACL enhances model generalization by maximizing attention consistency between local clients and the server model. To ensure privacy and validate robustness, we incorporated differential privacy by introducing noise during parameter transfer. We assessed the effectiveness of FACL in cancer diagnosis and Gleason grading tasks using 19,461 whole-slide images of prostate cancer from multiple centers. In the diagnosis task, FACL achieved an area under the curve (AUC) of 0.9718, outperforming seven centers with an average AUC of 0.9499 when categories are relatively balanced. For the Gleason grading task, FACL attained a Kappa score of 0.8463, surpassing the average Kappa score of 0.7379 from six centers. In conclusion, FACL offers a robust, accurate, and cost-effective AI training model for prostate cancer pathology while maintaining effective data safeguards.

Abstract Threshold voltage ( $$V_{th}$$ V th ) assignment is convenient for leakage optimization due to the exponential relation between leakage power and $$V_{th}$$ V th by swapping logic cells without routing effort. However, it poses great challenge in large scale circuit design as an NP-hard problem. Machine learning-based approaches have been proposed to solve this problem, aiming to achieve well tradeoff between leakage power reduction and runtime speed up without new induced timing violation. In this paper, a leakage power optimization framework based on reinforcement learning (RL) with graph neural network (GNN) is first-ever proposed to formulate $$V_{th}$$ V th assignment as a RL process by learning timing and physical characteristics of each circuit instance with GNN. Multiple instances are selected in a non-overlapped manner for each RL action iteration to speed up convergence and decouple timing interdependence along circuit path, where the corresponding reward is carefully defined to tradeoff between leakage reduction and potential timing violation. The proposed framework was validated by the Opencores and IWLS 2005 benchmark circuits with TSMC 28 nm technology. Experimental results demonstrate that our work outperforms prior non-analytical and GNN-based methods with better leakage power optimization by additional 5% to 17% reduction, which is highly consistent with the commercial tool. When transferring the trained RL-based framework to unseen circuits, it achieves the roughly identical leakage optimization results as seen circuit and speed up the runtime by 5.7 $$\times$$ × to 8.5 $$\times$$ × compared with commercial tool.

Background: Tendon-bone interface (TBI) repair is slow and challenging owing to its hierarchical structure, gradient composition, and complex function. In this work, enlightened by the natural characteristics of TBI microstructure and the demands of TBI regeneration, a structure, composition, and function-based scaffold was fabricated. Methods: The biomimetic scaffold was designed based on the “tissue-inducing biomaterials” theory: (1) a porous scaffold was created with poly-lactic-co-glycolic-acid, nano-hydroxyapatite and loaded with BMP2-gelatinmp to simulate the bone (BP); (2) a hydrogel was produced from sodium alginate, type I collagen, and loaded with TGF-β3 to simulate the cartilage (CP); (3) the L-poly-lactic-acid fibers were oriented to simulate the tendon (TP). The morphology of tri-layered constructs, gelation kinetics, degradation rate, release kinetics and mechanical strength of the scaffold were characterized. Then, bone marrow mesenchymal stem cells (MSCs) and tenocytes (TT-D6) were cultured on the scaffold to evaluate its gradient differentiation inductivity. A rat Achilles tendon defect model was established, and BMSCs seeded on scaffolds were implanted into the lesionsite. The tendon-bone lesionsite of calcaneus at 4w and 8w post-operation were obtained for gross observation, radiological evaluation, biomechanical and histological assessment. Results: The hierarchical microstructures not only endowed the scaffold with gradual composition and mechanical properties for matching the regional biophysical characteristics of TBI but also exhibited gradient differentiation inductivity through providing regional microenvironment for cells. Moreover, the scaffold seeded with cells could effectively accelerate healing in rat Achilles tendon defects, attributable to its enhanced differentiation performance. Conclusion: The hierarchical scaffolds simulating the structural, compositional, and cellular heterogeneity of natural TBI tissue performed therapeutic effects on promoting regeneration of TBI and enhancing the healing quality of Achilles tendon. The translational potential of this article: The novel scaffold showed the great efficacy on tendon to bone healing by offering a structural and compositional microenvironment. The results meant that the hierarchical scaffold with BMSCs may have a great potential for clinical application.

Abstract Background The novel anti-HER2 antibody drug conjugates (ADCs) can effectively improve the long-term survival of patients with HER2-low expression breast cancer. However, pathological responses to neoadjuvant therapy (NAT) within HER2-low expression breast cancer, the relationship between pathological response and prognosis and the transformation of HER2 status are all now poorly understood. Methods The patients with HER2-0 and HER2-low expression breast cancer receiving NAT at Harbin Medical University Cancer Hospital between Jan. 2014 and Nov. 2018 were retrospectively explored. HER2 low expression refers to the IHC 1 + or 2 + and FISH negative. The Kappa test was utilized for analyzing the consistency rate of HER2 expression. To evaluate disease-free survival (DFS) and overall survival (OS), this research employed both the Kaplan-Meier analysis and the Cox regression. Results In this study, 178 patients with HER2-0 and 344 patients with HER2-low expression breast cancer were included. In comparison with the HER2-0 group, it is shown that patients in the HER2-low group have more possibility to be younger compared to those 50 years old (P

Abstract The accuracy of short‐term photovoltaic (PV) power prediction is crucial for maintaining power system stability and grid scheduling. Here, a short‐term PV power prediction framework is proposed considering combined weather similarity day screening, signal decomposition noise reduction and hybrid deep learning to realize PV power prediction. First, a combined meteorological similar day screening model is constructed to screen out historical days similar to the day, which reduces the number of training sets; Second, Synchrosqueezing Wavelet Transform is utilized to eliminate data noise. Third, a Convolution Neural Network‐Gate Recurrent Unit (CNN‐GRU) network is constructed to extract periodic and nonlinear features in the PV power generation data series and to capture the relationship features between PV power generation and meteorological factors to improve the prediction accuracy. Fourth, the Sparrow Search Algorithm is introduced to perform hyper‐parameter optimization of the CNN‐GRU network to accelerate the model convergence and improve the model training efficiency. Finally, this paper conducts simulation experiments and the experimental results show that the prediction method proposed in this paper can effectively improve the prediction accuracy of short‐term PV power compared to the baseline model, and the method proposed in this paper is superior to other conventional methods.

Abstract Background and purpose The enzyme methylenetetrahydrofolate reductase (MTHFR) plays a crucial role in directing folate species towards nucleotide synthesis or DNA methylation. The MTHFR polymorphisms C677T and A1298C have been linked to cancer susceptibility, but the evidence supporting this association has been equivocal. To investigate the individual and joint associations between MTHFR C677T, A1298C, and digestive system cancer in a Chinese hypertensive population, we conducted a population-based case–control study involving 751 digestive system cancer cases and one-to-one matched controls from the China H-type Hypertension Registry Study (CHHRS). Methods We utilized the conditional logistic regression model to evaluate multivariate odds ratios (ORs) and 95% confidence intervals (CIs) of digestive system cancer. Results The analysis revealed a significantly lower risk of digestive system cancer in individuals with the CT genotype (adjusted OR: 0.71; 95% CI 0.52, 0.97; P = 0.034) and TT genotype (adjusted OR: 0.57; 95% CI 0.40, 0.82; P = 0.003; P for trend = 0.003) compared to those with the 677CC genotype. Although A1298C did not show a measurable association with digestive system cancer risk, further stratification of 677CT genotype carriers by A1298C homozygotes (AA) and heterozygotes (AC) revealed a distinct trend within these subgroups. Conclusion These findings indicate a potential protective effect against digestive system cancer associated with the T allele of MTHFR C677T. Moreover, we observed that the presence of different combinations of MTHFR polymorphisms may contribute to varying susceptibilities to digestive system cancer.

BackgroundRecent studies have shown that gut microbiota may be related to the occurrence of brain tumors, but direct evidence is lacking. This study used the Mendelian randomization study (MR) method to explore the potential causal link between gut microbiota and brain tumors.MethodWe analyzed the genome-wide association data between 211 gut microbiota taxa and brain tumors, using the largest existing gut microbiota Genome-Wide Association Studies meta-analysis data (n=13266) and combining it with brain tumor data in the IEU OpenGWAS database. We use inverse-variance weighted analysis, supplemented by methods such as Mendelian randomization-Egger regression, weighted median estimator, simple mode, and weighted mode, to assess causality. In addition, we also conducted the Mendelian randomization-Egger intercept test, Cochran’s Q test, and Mendelian randomization Steiger directionality test to ensure the accuracy of the analysis. Quality control includes sensitivity analysis, horizontal gene pleiotropy test, heterogeneity test, and MR Steiger directionality test.ResultOur study found that specific gut microbial taxa, such as order Lactobacillales and family Clostridiaceae1, were positively correlated with the occurrence of brain tumors, while genus Defluviitaleaceae UCG011 and genus Flavonifractor were negatively correlated with the occurrence of brain tumors. The Mendelian randomization-Egger intercept test showed that our analysis was not affected by pleiotropy (P>0.05).ConclusionThis study reveals for the first time the potential causal relationship between gut microbiota and brain tumors, providing a new perspective for the prevention and treatment of early brain tumors. These findings may help develop new clinical intervention strategies and point the way for future research.

As a promising alternative to traditional underwater communication methods, underwater wireless optical communication (UWOC) offers higher data rates, lower latency, and enhanced security. However, the practical implementation of high-speed UWOC systems typically faces challenges arising from bandwidth-limited optical devices and adverse effects of absorption, scattering, and oceanic turbulence. To tackle these challenges, we propose a novel joint probabilistic shaping and precoding scheme that simultaneously optimizes the mutual information (MI) and reduces the peak-to-average power ratio (PAPR) of multiple-input single-output (MISO) orthogonal frequency-division multiplexing (OFDM) UWOC systems. Specifically, the proposed approach probabilistically shapes the transmitted symbols and appropriately allocates weights to both multi-light sources and multi-carriers based on precoding to match the channel characteristics. Numerical results demonstrate its superiority over existing schemes, particularly showcasing its potential to significantly improve the performance of UWOC systems via MI optimization and PAPR reduction. Moreover, the results verify that the proposed approach exhibits satisfactory robustness against oceanic turbulence.

Abstract Background NONO-TFE3 rearranged renal cell carcinoma (NONO-TFE3 rRCC) is one of a subtype of TFE3 rRCCs with high malignancy and poor prognosis. Compared with clear cell RCC, NONO-TFE3 rRCC shows a preference for mitochondrial respiration. We recently identified that the upregulation of nicotinamide ribokinase 2 (NMRK2) was associated with enhanced mitochondrial respiration and tumor progression in TFE3 rRCC. Methods A tumor-bearing mouse model was established to verify the pro-oncogenic effect of NMRK2 on NONO-TFE3 rRCC. Then the expression of NMRK2 RNA and protein was detected in cell lines and patient specimens. The NMRK2 transcripts were Sanger-sequenced and blasted at NCBI website. We constructed dCas13b-HA system to investigate the factors binding with NMRK2 RNA. We also used molecular experiments like RIP-seq, IP-MS, FISH and fluorescence techniques to explore the mechanisms that long non-coding RNA (lncRNA) like NMRK2 mRNA promoted the mitochondrial respiration of NONO-TFE3 rRCC. The efficacy of the combination of shRNA (NMRK2)-lentivirus and metformin on NONO-TFE3 rRCC was assessed by CCK-8 assay. Results In this study, we confirmed that NMRK2 showed transcriptional-translational conflict and functioned as lncRNA like mRNA in the NONO-TFE3 rRCC. Furthermore, we revealed the molecular mechanism that NONO-TFE3 fusion suppressed the translation of NMRK2 mRNA. Most importantly, three major pathways were shown to explain the facilitation effects of lncRNA like NMRK2 mRNA on the mitochondrial respiration of NONO-TFE3 rRCC in an NAD+ kinase-independent manner. Finally, the efficacy of combination of shRNA (NMRK2)-lentivirus and metformin on NONO-TFE3 rRCC was demonstrated to be superior than either agent alone. Conclusions Overall, our data comprehensively demonstrated the mechanisms for the enhanced mitochondrial respiration in NONO-TFE3 rRCC and proposed lncRNA like NMRK2 mRNA as a therapy target for NONO-TFE3 rRCC. Graphical Abstract

As a key submarine-based communication platform, unmanned underwater vehicle (UUV) can facilitate underwater wireless optical communication (UWOC).However, fluctuating characteristics of water body, different water qualities, multi-user access present challenges to UUV-assisted UWOC systems, which could be alleviated by an appropriate path planning to maximize the system and each user performance.Deep reinforcement learning (DRL) technology was applied in the path planning of autonomous vehicles, a trajectory planning scheme for UUV-assisted UWOC systems was proposed.The UUV automatically decides the navigation direction through deep Q-network (DQN) method, thereby improving the communication capacity of the system and each user.The impact of distinct water qualities on the capacity enhancement was also investigated.Simulation results suggest that the outputted strategy of DQN can improve the link capacity of the system and each user.This capacity improvement in clear seawater is better than that in pure seawater but lower than that in coastal water.

Underwater wireless optical communication (UWOC) has been widely used in the construction of underwater communication network for the advantages of high data rates, small latency, high flexibility and scalability.However, due to the absorption and scattering effects of seawater, the spatial characteristics of UWOC links have not been studied extensively, especially in multiple-input multiple-output (MIMO) scenario.A linear array MIMO UWOC link geometry was considered and a weighted exponential function polynomial (WEFP) model was proposed to characterize the photon irradiance distribution at the receiving plane of general MIMO UWOC links with arbitrary numbers of light sources.Numerical results suggest that the proposed WEFP model fits well with the results of Monte Carlo simulations in turbid water environment.

Objective: Bone regeneration involves a coordinated cascade of events that are regulated by several cytokines and growth factors, among which bone morphogenic protein-2 (BMP-2), vascular endothelial growth factor (VEGF) and fibroblast growth factor-2 (FGF-2) play important roles. In this study, we investigated the effects of dual release of the three growth factors on bone regeneration in femur defects. Methods: A composite consisting of Gelatin microparticles loaded with VEGF/FGF-2 and poly(lactic-co-glycolic acid)-poly(ethylene glycol)-carboxyl (PLGA-PEG-COOH) microparticles loaded with BMP-2 encapsulated in a nano hydroxyapatite-poly actic-co-glycolic acid (nHA-PLGA) scaffold was prepared for the dual release of the growth factors. Results: On the 14th day, decreased release rate of BMP-2 compared with FGF-2 and VEGF was observed. However, after 14 days, compared to FGF-2 and VEGF, BMP-2 showed an increased release rate. Controlled dual release of BMP-2 and VEGF, FGF-2 resulted in a significant osteogenic differentiation of bone mesenchymal stem cells (BMSCs). Moreover, effects of the composite scaffold on functional connection of osteoblast-vascular cells during bone development were evaluated. The synergistic effects of dual delivery of growth factors were shown to promote the expression of VEGF in BMSCs. Increased secretion of VEGF from BMSCs promoted the proliferation and angiogenic differentiation of human umbilical vein endothelial cells (HUVECs) in the co-culture system. At 12 weeks after implantation, blood vessel and bone formation were analyzed by micro-CT and histology. The composite scaffold significantly promoted the formation of blood vessels and new bone in femur defects. Conclusions: These findings demonstrate that dual delivery of angiogenic factors and osteogenic factors from Gelatin and PLGA-PEG-COOH microparticles-based composite scaffolds exerted an osteogenic-angiogenic coupling effect on bone regeneration. This approach will inform on the development of appropriate designs of high-performance bioscaffolds for bone tissue engineering.

Deep learning techniques have shown great potential in medical image processing, particularly through accurate and reliable image segmentation on magnetic resonance imaging (MRI) scans or computed tomography (CT) scans, which allow the localization and diagnosis of lesions. However, training these segmentation models requires a large number of manually annotated pixel-level labels, which are time-consuming and labor-intensive, in contrast to image-level labels that are easier to obtain. It is imperative to resolve this problem through weakly-supervised semantic segmentation models using image-level labels as supervision since it can significantly reduce human annotation efforts. Most of the advanced solutions exploit class activation mapping (CAM). However, the original CAMs rarely capture the precise boundaries of lesions. In this study, we propose the strategy of multi-scale inference to refine CAMs by reducing the detail loss in single-scale reasoning. For segmentation, we develop a novel model named Mixed-UNet, which has two parallel branches in the decoding phase. The results can be obtained after fusing the extracted features from two branches. We evaluate the designed Mixed-UNet against several prevalent deep learning-based segmentation approaches on our dataset collected from the local hospital and public datasets. The validation results demonstrate that our model surpasses available methods under the same supervision level in the segmentation of various lesions from brain imaging.

With the development of light emitting diode (LED) lighting technology and its wide applications, visible light communication (VLC) technology has also seen significant advancements. VLC is regarded as a supplementary technology to radio frequency (RF) due to its unregulated spectrum and extraordinarily high communication rates. In this paper, the advantages, architectures, key technologies, application scenarios and machine learning (ML) applications of VLC are reviewed and summarized.

Recently, integrated science, technology, engineering, and mathematics (STEM) education has gained sustained attention in K-12 settings, and engineering design-based pedagogy has become a key issue. Compared with rich research in higher education, relatively few studies are performed on engineering education in K-12 schools. In this study, we combined Conceive-Design-Implement-Operate (CDIO) model with the engineering design process (EDP), naming EDP-CDIO, aiming to promote high school students' STEM competence and compare its effects with the conventional CDIO approach. A pretest--posttest nonequivalent group design was conducted among 64 eleventh-grade students with eleven lessons. Quantitative data were collected via a pretest and posttest, and qualitative data were collected via artifacts and semistructured interviews. The repeated-measures analysis of variance and epistemic network analysis revealed that, compared with the conventional CDIO approach, the EDP-CDIO model significantly improved students' STEM knowledge, skills, and attitudes and developed more comprehensive epistemic networks in STEM competence. These findings provide a reference for K-12 STEM teachers, encouraging them to implement the EDP-CDIO model more frequently in the classroom, especially with the iterative design process.

An wireless catching and transferring system of next generation network was introduced. With the advanced technology of IPv6, IEEE 802.11n and cloud computing, the system provided an effective solution, meeting the needs of professional journalists to catch and transfer quickly at large games and meetings.

The miniaturization of spectrometer can broaden the application area of spectrometry, which has huge academic and industrial value. Among various miniaturization approaches, filter-based miniaturization is a promising implementation by utilizing broadband filters with distinct transmission functions. Mathematically, filter-based spectral reconstruction can be modeled as solving a system of linear equations. In this paper, we propose an algorithm of spectral reconstruction based on sparse optimization and dictionary learning. To verify the feasibility of the reconstruction algorithm, we design and implement a simple prototype of a filter-based miniature spectrometer. The experimental results demonstrate that sparse optimization is well applicable to spectral reconstruction whether the spectra are directly sparse or not. As for the non-directly sparse spectra, their sparsity can be enhanced by dictionary learning. In conclusion, the proposed approach has a bright application prospect in fabricating a practical miniature spectrometer.

In this article, a novel driving behavior recognition system based on a specific physical model and motion sensory data is developed to promote traffic safety. Based on the theory of rigid body kinematics, we build a specific physical model to reveal the data change rule during the vehicle moving process. In this work, we adopt a nine-axis motion sensor including a three-axis accelerometer, a three-axis gyroscope and a three-axis magnetometer, and apply a Kalman filter for noise elimination and an adaptive time window for data extraction. Based on the feature extraction guided by the built physical model, various classifiers are accomplished to recognize different driving behaviors. Leveraging the system, normal driving behaviors (such as accelerating, braking, lane changing and turning with caution) and aggressive driving behaviors (such as accelerating, braking, lane changing and turning with a sudden) can be classified with a high accuracy of 93.25%. Compared with traditional driving behavior recognition methods using machine learning only, the proposed system possesses a solid theoretical basis, performs better and has good prospects.