Adaptive Decision Spatio-temporal neural ODE for traffic flow forecasting with Multi-Kernel Temporal Dynamic Dilation Convolution.

Traffic flow prediction is crucial for efficient traffic management. It involves predicting vehicle movement patterns to reduce congestion and enhance traffic flow. However, the highly non-linear and complex patterns commonly observed in traffic flow pose significant challenges for this task. Current Graph Neural Network (GNN) models often construct shallow networks, which limits their ability to extract deeper spatio-temporal representations. Neural ordinary differential equations for traffic prediction address over-smoothing but require significant computational resources, leading to inefficiencies, and sometimes deeper networks may lead to poorer predictions for complex traffic information. In this study, we propose an Adaptive Decision spatio-temporal Neural Ordinary Differential Network, which can adaptively determine the number of layers of ODE according to the complexity of traffic information. It can solve the over-smoothing problem better, improving overall efficiency and prediction accuracy. In addition, traditional temporal convolution methods make it difficult to deal with complex and variable traffic time information with a large time span. Therefore, we introduce a multi-kernel temporal dynamic expansive convolution to handle the traffic time information. Multi-kernel temporal dynamic expansive convolution employs a dynamic dilation strategy, dynamically adjusting the network's receptive field across levels, effectively capturing temporal dependencies, and can better adapt to the changing time data of traffic information. Additionally, multi-kernel temporal dynamic expansive convolution integrates multi-scale convolution kernels, enabling the model to learn features across diverse temporal scales. We evaluated our proposed method on several real-world traffic datasets. Experimental results show that our method outperformed state-of-the-art benchmarks.

A near-resonant excitation strategy to achieve ultra-low threshold GaN polariton lasing.

A near-resonant excitation strategy is proposed and implemented in a 4-µm-thick GaN microcavity to realize an exciton-polariton condensate/lasing with low threshold. Strong exciton-photon coupling is demonstrated, and polariton lasing is realized with an ultra-low threshold excitation power density of about 13.3 W/cm2 at room temperature. Such an ultra-low threshold is ascribed to the implementation of the near-resonant optical excitation strategy, which enables acceleration of the exciton and polariton relaxation and suppression of the heat generation in the cavity, thereby reducing the energy loss and enhance the cavity excitation efficiency.

Nutrition-Related Knowledge Graph Neural Network for Food Recommendation.

Food recommendation systems are becoming increasingly vital in modern society, given the fast-paced lifestyle and diverse dietary habits. Existing research and implemented solutions often rely on user preferences and past behaviors for recommendations, which poses significant issues. Firstly, this approach inadequately considers the nutritional content of foods, potentially leading to recommendations that are overly homogeneous and lacking in diversity. Secondly, it may result in repetitive suggestions of the same types of foods, thereby encouraging users to develop unhealthy dietary habits that could adversely affect their overall health. To address this issue, we introduce a novel nutrition-related knowledge graph (NRKG) method based on graph convolutional networks (GCNs). This method not only enhances users' ability to select appropriate foods but also encourages the development of healthy eating habits, thereby contributing to overall public health. The NRKG method comprises two key components: user nutrition-related food preferences and recipe nutrition components. The first component gathers nutritional information from recipes that users show interest in and synthesizes these data for user reference. The second component connects recipes with similar nutritional profiles, forming a complex heterogeneous graph structure. By learning from this graph, the NRKG method integrates user preferences with nutritional data, resulting in more accurate and personalized food recommendations. We evaluated the NRKG method against six baseline methods using real-world food datasets. In the 100% dataset, the five metrics exceeded the performance of the best baseline method by 2.8%, 5.9%, 1.5%, 9.7%, and 6.0%, respectively. The results indicate that our NRKG method significantly outperforms the baseline methods, including FeaStNet, DeepGCN, GraphSAGE, GAT, UniMP, and GATv2, demonstrating its superiority and effectiveness in promoting healthier and more diverse eating habits. Unlike these baseline methods, which primarily focus on hierarchical information propagation, our NRKG method offers a more comprehensive approach by integrating the nutritional information of recipes with user preferences.

Smart Biosensor for Breast Cancer Survival Prediction Based on Multi-View Multi-Way Graph Learning.

Biosensors play a crucial role in detecting cancer signals by orchestrating a series of intricate biological and physical transduction processes. Among various cancers, breast cancer stands out due to its genetic underpinnings, which trigger uncontrolled cell proliferation, predominantly impacting women, and resulting in significant mortality rates. The utilization of biosensors in predicting survival time becomes paramount in formulating an optimal treatment strategy. However, conventional biosensors employing traditional machine learning methods encounter challenges in preprocessing features for the learning task. Despite the potential of deep learning techniques to automatically extract useful features, they often struggle to effectively leverage the intricate relationships between features and instances. To address this challenge, our study proposes a novel smart biosensor architecture that integrates a multi-view multi-way graph learning (MVMWGL) approach for predicting breast cancer survival time. This innovative approach enables the assimilation of insights from gene interactions and biosensor similarities. By leveraging real-world data, we conducted comprehensive evaluations, and our experimental results unequivocally demonstrate the superiority of the MVMWGL approach over existing methods.

Perovskite Quantum Dots Lasing in Double-Heterostructure through Energy Transfer.

Planar double heterostructures were initially investigated and have been successfully applied in III-V semiconductor lasers due to their excellent roles in confining both the photons and carriers. Here, we design and fabricate a (PEA)2Csn-1PbnX3n+1 (quasi-2D)/CsPbBr3 QD/quasi-2D double-heterostructure sandwiched in a 3/2 λ DBR microcavity, and then demonstrate a single-mode pure-green lasing with a threshold of 53.7 µJ/cm2 under nanosecond-pulsed optical pumping. The thresholds of these heterostructure devices decrease statistically by about 50% compared to the control group with no energy donor layers, PMMA/QD/PMMA in an identical microcavity. We show that there is efficient energy transfer from the barrier regions of the quasi-2D phases to the QD layer by transient absorption and luminescence lifetime spectra and that such energy transfer leads to marked threshold reduction. This work indicates that the double-heterostructure configurations should play a significant role in the future perovskite electrically pumped laser.

Regulating Alkyl Chain Length on Quaternization TiO2 for Boosting Photocatalytic Performance: Synergism of Promoting Photogenerated Charge Separation and Improving Reactant Adsorption.

To explore the effect of surface charge properties of a photocatalyst on photocatalytic activity, quaternization TiO2 particles with different alkyl chain lengths (CT-X/P25) are first synthesized to boost the photocatalytic activity. The effect of a quaternary ammonium group with different alkyl chain lengths on the photocatalytic activity of CT-X/P25 is investigated. Interestingly, the introduction of a quaternary ammonium group on CT-X/P25 not only contributes to improving the photodegradation efficiency of anionic dyes due to enhancing of adsorption capacity through electrostatic attraction, but also it can improve the photodegradation efficiency of cationic dyes. Therefore, there seems to be another factor affecting the photocatalytic activity. The results of photoelectric characterization show that the photogenerated charge separation of CT-X/P25 is greatly enhanced, which is beneficial to improve the photocatalytic activity. Simultaneously, the results show that the difference in the photocatalytic activity of CT-X/P25 is mainly related to the charge intensity of -N+(CH3)2- in the quaternary amine salt. According to X-ray photoelectron spectroscopy, the charge intensity of -N+(CH3)2- in CT-X/P25 gradually increases with the increase in alkyl chain lengths, which is conducive to promoting photogenerated charge separation and improving the adsorption for anionic dyes. The photocatalytic activity has been further enhanced due to the enhancement of this synergy. In summary, the quaternary ammonium salt-modified CT-X/P25 shows an excellent synergistic effect on the process of photodegradation of anionic dyes: promoting photogenerated charge separation and adsorption.

Large-area large-grain CsPbCl3perovskite films by confined re-growth for violet photodetectors.

CsPbCl3perovskite is an attractive semiconductor material with characteristics such as a wide bandgap, high chemical stability, and excellent optoelectronic properties, which broaden its application prospects for ultraviolet (UV) and violet photodetectors (PDs). However, large-area CsPbCl3films with high coverage, large grains, and controllable thickness are still difficult to prepare by using the solution method due to the extremely low solubility of their precursors in conventional solvents. Herein, a water-assisted confined re-growth method is developed, and a CsPbCl3microcrystalline film with an area of 3 cm × 3 cm is grown, the thickness of which is controllable within a range of several microns. The as-prepared thin film exhibits a flat and smooth surface, large grains, and enhanced photoluminescence. Furthermore, the fabricated violet PDs based on the prepared CsPbCl3film show a high responsivity of 2.17 A W-1, external quantum efficiency of 664%, on/off ratio of 2.58 × 103, and good stability. This study provides a prospective solution for the growth of large-area, large-grain, and surface-smooth CsPbCl3films for high-performance UV and violet PDs.

Growth of hopper-shaped CsPbCl3 crystals and their exciton polariton emission.

CsPbCl3 is an attractive wide-bandgap perovskite semiconductor. Herein, we have grown hopper-shaped CsPbCl3 crystals in a solution droplet dripped on a heated substrate. During the growth, we have observed the impacts of the coffee ring effect and Marangoni flow, which may result in the hopper shape. Their photoluminescence spectra feature double peaks, which are located at 413.9 nm and 422.0 nm, respectively, and the latter increases faster in intensity than the former as the excitation power increases. We believe that the higher-energy peak originates from the excitonic emission and the lower-energy one is from the polaritons' emission, where the polaritons are generated in the exciton-exciton inelastic scattering process. Based on such an explanation, the exciton binding energy of CsPbCl3 is estimated to be 76.7 meV in our experiments, consistent with the previous reports.