The parents' internet use and children's extracurricular tutoring class participation.

The educational burden from extracurricular tutoring class has become a pressing social issue in China. This study used data from the China family panel studies (CFPS) in 2014, 2016, and 2018 to empirically analyze the impact of Internet usage on children's participation in extracurricular tutoring class. There are many factors that influence parents' decisions to enroll their children in extracurricular tutoring class. These factors include family income status, the level of importance parents place on their children's education, the marginal returns on educational investment, academic pressure, etc. However, in today's digitalized society, the widespread use of the internet will also become an important influencing factor in parents' decisions regarding educational investment. The study finds that, parents by using the Internet significantly increase the probability of enrolling their children in extracurricular tutoring class. Through mechanism regression analysis, it is concluded that internet usage has a positive influence on parents enrolling their children in extracurricular tutoring class by increasing the frequency of social interaction and raising parents' educational expectations for their children. Based on the empirical results, the following policy suggestions were proposed: 1. Schools should establish a more comprehensive after-school education service system to improve the engagement of students in compulsory education; 2. The government can enhance the accessibility and optimization of educational resources by increasing investment in education, improving the quality of in-school education, and optimizing the management and supervision of extracurricular tutoring class. This ensures that students can access high-quality educational services.

Exogenous 24-Epibrassinolide Enhanced Drought Tolerance and Promoted BRASSINOSTEROID-INSENSITIVE2 Expression of Quinoa.

Brassinosteroids (BRs) are involved in the regulation of biotic and abiotic stresses in plants. The molecular mechanisms of BRs that alleviate the drought stress in quinoa have rarely been reported. Here, quinoa seedlings were treated with 24-epibrassinolide (EBR) and we transiently transferred CqBIN2 to the quinoa seedlings' leaves using VIGS technology to analyze the molecular mechanism of the BR mitigation drought stress. The results showed that EBR treatment significantly increased the root growth parameters, the antioxidant enzyme activities, and the osmolyte content, resulting in a decrease in the H2O2, O2∙-, and malondialdehyde content in quinoa. A transcriptome analysis identified 8124, 2761, and 5448 differentially expressed genes (DEGs) among CK and Drought, CK and EBR + Drought, and Drought and EBR + Drought groups. WGCNA divided these DEGs into 19 modules in which these characterized genes collectively contributed significantly to drought stress. In addition, the EBR application also up-regulated the transcript levels of CqBIN2 and proline biosynthesis genes. Silenced CqBIN2 by VIGS could reduce the drought tolerance, survival rate, and proline content in quinoa seedlings. These findings not only revealed that exogenous BRs enhance drought tolerance, but also provided insight into the novel functions of CqBIN2 involved in regulating drought tolerance in plants.

Optical Second Harmonic Generation of Low-Dimensional Semiconductor Materials.

In recent years, the phenomenon of optical second harmonic generation (SHG) has attracted significant attention as a pivotal nonlinear optical effect in research. Notably, in low-dimensional materials (LDMs), SHG detection has become an instrumental tool for elucidating nonlinear optical properties due to their pronounced second-order susceptibility and distinct electronic structure. This review offers an exhaustive overview of the generation process and experimental configurations for SHG in such materials. It underscores the latest advancements in harnessing SHG as a sensitive probe for investigating the nonlinear optical attributes of these materials, with a particular focus on its pivotal role in unveiling electronic structures, bandgap characteristics, and crystal symmetry. By analyzing SHG signals, researchers can glean invaluable insights into the microscopic properties of these materials. Furthermore, this paper delves into the applications of optical SHG in imaging and time-resolved experiments. Finally, future directions and challenges toward the improvement in the NLO in LDMs are discussed to provide an outlook in this rapidly developing field, offering crucial perspectives for the design and optimization of pertinent devices.

Galactinol synthase 2 influences the metabolism of chlorophyll, carotenoids, and ethylene in tomato fruits.

Galactinol synthase (GolS), which catalyses the synthesis of galactinol, is the first critical enzyme in the biosynthesis of raffinose family oligosaccharides (RFOs) and contributes to plant growth and development, and resistance mechanisms. However, its role in fruit development remains largely unknown. In this study, we used CRISPR/Cas9 gene-editing technology in tomato (Solanum lycopersicum) to create the gols2 mutant showing uniformly green fruits without dark-green shoulders, and promoting fruit ripening. Analysis indicated that galactinol was undetectable in the ovaries and fruits of the mutant, and the accumulation of chlorophyll and chloroplast development was suppressed in the fruits. RNA-sequencing analysis showed that genes related to chlorophyll accumulation and chloroplast development were down-regulated, including PROTOCHLOROPHYLLIDE OXIDOREDUCTASE, GOLDEN 2-LIKE 2, and CHLOROPHYLL A/B-BINDING PROTEINS. In addition, early color transformation and ethylene release was prompted in the gols2 lines by regulation of the expression of genes involved in carotenoid and ethylene metabolism (e.g. PHYTOENE SYNTHASE 1, CAROTENE CIS-TRANS ISOMERASE, and 1-AMINOCYCLOPROPANE-1-CARBOXYLIC ACID SYNTHASE2/4) and fruit ripening (e.g. RIPENING INHIBITOR, NON-RIPENING, and APETALA2a). Our results provide evidence for the involvement of GolS2 in pigment and ethylene metabolism of tomato fruits.

Single-cell profiling of African swine fever virus disease in the pig spleen reveals viral and host dynamics.

African swine fever, one of the major viral diseases of swine, poses an imminent threat to the global pig industry. The high-efficient replication of the causative agent African swine fever virus (ASFV) in various organs in pigs greatly contributes to the disease. However, how ASFV manipulates the cell population to drive high-efficient replication of the virus in vivo remains unclear. Here, we found that the spleen reveals the most severe pathological manifestation with the highest viral loads among various organs in pigs during ASFV infection. By using single-cell-RNA-sequencing technology and multiple methods, we determined that macrophages and monocytes are the major cell types infected by ASFV in the spleen, showing high viral-load heterogeneity. A rare subpopulation of immature monocytes represents the major population infected at late infection stage. ASFV causes massive death of macrophages, but shifts its infection into these monocytes which significantly arise after the infection. The apoptosis, interferon response, and antigen-presentation capacity are inhibited in these monocytes which benefits prolonged infection of ASFV in vivo. Until now, the role of immature monocytes as an important target by ASFV has been overlooked due to that they do not express classical monocyte marker CD14. The present study indicates that the shift of viral infection from macrophages to the immature monocytes is critical for maintaining prolonged ASFV infection in vivo. This study sheds light on ASFV tropism, replication, and infection dynamics, and elicited immune response, which may instruct future research on antiviral strategies.

Integrating visual large language model and reasoning chain for driver behavior analysis and risk assessment.

Driver behavior is a critical factor in driving safety, making the development of sophisticated distraction classification methods essential. Our study presents a Distracted Driving Classification (DDC) approach utilizing a visual Large Language Model (LLM), named the Distracted Driving Language Model (DDLM). The DDLM introduces whole-body human pose estimation to isolate and analyze key postural features-head, right hand, and left hand-for precise behavior classification and better interpretability. Recognizing the inherent limitations of LLMs, particularly their lack of logical reasoning abilities, we have integrated a reasoning chain framework within the DDLM, allowing it to generate clear, reasoned explanations for its assessments. Tailored specifically with relevant data, the DDLM demonstrates enhanced performance, providing detailed, context-aware evaluations of driver behaviors and corresponding risk levels. Notably outperforming standard models in both zero-shot and few-shot learning scenarios, as evidenced by tests on the 100-Driver dataset, the DDLM stands out as an advanced tool that promises significant contributions to driving safety by accurately detecting and analyzing driving distractions.

FGL1 in plasma extracellular vesicles is correlated with clinical stage of lung adenocarcinoma and anti-PD-L1 response.

Fibrinogen-like protein-1 (FGL1) is confirmed a major ligand of lymphocyte activation gene-3 which could inhibit antigen-mediated T-cell response and evade immune supervision. Although hepatocytes secrete large amounts of FGL1, its high expression also be detected in solid tumors such as lung cancer, leading to a poor efficacy of immune checkpoint inhibitors therapy. Here we reported that FGL1 was overexpressed in lung adenocarcinoma (LUAD) but not in lung squamous cell carcinoma. However, FGL1 in tissue and plasma can only distinguish LUAD patients from healthy donors and cannot correlate with clinical Tumor Node Metastasis (TNM) stage. Using lung cancer cell lines, we confirmed that FGL1 can be detected on extracellular vesicles (EVs) and we established a method using flow cytometry to detect FGL1 on the surface of EVs, which revealed that FGL1 could be secreted via EVs. Both animal model and clinical samples proved that plasma FGL1 in EVs would increase when the tumor was loaded. The level of FGL1 in plasma EVs was correlated with clinical TNM stage and tumor size, and a higher level indicated non-responsiveness to anti-programmed cell death ligand 1 (anti-PD-L1) immunotherapy. Its effect on tumor progression and immune evasion may be achieved by impairing the killing and proliferating capacities of CD8+ T cells. Our result demonstrates that FGL1 levels in plasma EVs, but not total plasma FGL1, could be a promising biomarker that plays an important role in predicting anti-PD-L1 immune therapy in LUAD and suggests a new strategy in LUAD immunotherapy.

Unlocking the double-dividend: Evaluating the impact of SO2 emissions trading scheme on firm's environmental and economic performance.

The optimal design of environmental instruments demands a balance between environmental enhancement and economic growth. Utilizing microdata from the China Environmental Statistics Database and the China Industrial Firm Database, this study employs the difference-in-differences (DD) methodology to explore the dual effects of the SO2 Emissions Trading Scheme (ETS) on the environmental and economic performance of micro-firms. The findings suggest that: (1) The SO2 ETS not only induces emission reduction effects among firms in pilot areas but also improves their industrial added value. (2) The SO2 ETS exhibits heterogeneous impacts across firms of diverse ownership, export status, and size. (3) While the SO2 ETS prompts firms to advance technologically, boosting desulfurization capacities and subsequently enhancing total factor productivity, it also inadvertently results in companies offsetting some environmental compliance costs by curtailing employee wages.

Tomato short internodes and pedicels encode an LRR receptor-like serine/threonine-protein kinase ERECTA regulating stem elongation through modulating gibberellin metabolism.

Plant height is an important agronomic trait. Dwarf varieties present several advantages, such as lodging resistance, increased yield, and suitability for mechanized harvesting, which are crucial for crop improvement. However, limited research is available on dwarf tomato varieties suitable for production. In this study, we report a novel short internode mutant named "short internode and pedicel (sip)" in tomato, which exhibits marked internode and pedicel shortening due to suppressed cell elongation. This mutant plant has a compact plant structure and compact inflorescence, and has been demonstrated to produce more fruits, resulting in a higher harvest index. Genetic analysis revealed that this phenotype is controlled by a single recessive gene, SlSIP. BSA analysis and KASP genotyping indicated that ERECTA (ER) is the possible candidate gene for SlSIP, which encodes a leucine-rich receptor-like kinase. Additionally, we obtained an ER functional loss mutant using the CRISPR/Cas9 gene-editing technology. The 401st base A of ER is substituted with T in sip, resulting in a change in the 134th amino acid from asparagine (N) to isoleucine (I). Molecular dynamics(MD) simulations showed that this mutation site is located in the extracellular LRR domain and alters nearby ionic bonds, leading to a change in the spatial structure of this site. Transcriptome analysis indicated that the genes that were differentially expressed between sip and wild-type (WT) plants were enriched in the gibberellin metabolic pathway. We found that GA3 and GA4 decreased in the sip mutant, and exogenous GA3 restored the sip to the height of the WT plant. These findings reveal that SlSIP in tomatoes regulates stem elongation by regulating gibberellin metabolism. These results provide new insights into the mechanisms of tomato dwarfing and germplasm resources for breeding dwarfing tomatoes.

Transcriptome analysis of Artemisia argyi following methyl jasmonate (MeJA) treatment and the mining of genes related to the stress resistance pathway.

Artemisia argyi Lev. et Vant. (A. argyi) is a perennial grass in the Artemisia family, the plant has a strong aroma. Methyl jasmonate (MeJA) is critical to plant growth and development, stress response, and secondary metabolic processes. The experimental material Artemisia argyi was utilized in this study to investigate the treatment of A. argyi with exogenous MeJA at concentrations of 100 and 200 µmol/L for durations of 9 and 24 h respectively. Transcriptome sequencing was conducted using the Illumina HiSeq platform to identify stress resistance-related candidate genes. Finally, a total of 102.43 Gb of data were obtained and 162,272 unigenes were identified. Differential analysis before and after MeJA treatment resulted in the screening of 20,776 differentially expressed genes. The GO classification revealed that the annotated unigenes were categorized into three distinct groups: cellular component, molecular function, and biological process. Notably, binding, metabolic process, and cellular process emerged as the most prevalent categories among them. The results of KEGG pathway statistical analysis revealed that plant hormone signal transduction, MAPK signaling pathway-plant, and plant-pathogen interaction were significant transduction pathways in A. argyi's response to exogenous MeJA-induced abiotic stress. With the alteration of exogenous MeJA concentration and duration, a significant upregulation was observed in the expression levels of calmodulin CaM4 (ID: EVM0136224) involved in MAPK signaling pathway-plant and auxin response factor ARF (ID: EVM0055178) associated with plant-pathogen interaction. The findings of this study establish a solid theoretical foundation for the future development of highly resistant varieties of A. argyi.

IBGJO: Improved Binary Golden Jackal Optimization with Chaotic Tent Map and Cosine Similarity for Feature Selection.

Feature selection is a crucial process in machine learning and data mining that identifies the most pertinent and valuable features in a dataset. It enhances the efficacy and precision of predictive models by efficiently reducing the number of features. This reduction improves classification accuracy, lessens the computational burden, and enhances overall performance. This study proposes the improved binary golden jackal optimization (IBGJO) algorithm, an extension of the conventional golden jackal optimization (GJO) algorithm. IBGJO serves as a search strategy for wrapper-based feature selection. It comprises three key factors: a population initialization process with a chaotic tent map (CTM) mechanism that enhances exploitation abilities and guarantees population diversity, an adaptive position update mechanism using cosine similarity to prevent premature convergence, and a binary mechanism well-suited for binary feature selection problems. We evaluated IBGJO on 28 classical datasets from the UC Irvine Machine Learning Repository. The results show that the CTM mechanism and the position update strategy based on cosine similarity proposed in IBGJO can significantly improve the Rate of convergence of the conventional GJO algorithm, and the accuracy is also significantly better than other algorithms. Additionally, we evaluate the effectiveness and performance of the enhanced factors. Our empirical results show that the proposed CTM mechanism and the position update strategy based on cosine similarity can help the conventional GJO algorithm converge faster.

Intestinal microbiota links to allograft stability after lung transplantation: a prospective cohort study.

Whether the alternated microbiota in the gut contribute to the risk of allograft rejection (AR) and pulmonary infection (PI) in the setting of lung transplant recipients (LTRs) remains unexplored. A prospective multicenter cohort of LTRs was identified in the four lung transplant centers. Paired fecal and serum specimens were collected and divided into AR, PI, and event-free (EF) groups according to the diagnosis at sampling. Fecal samples were determined by metagenomic sequencing. And metabolites and cytokines were detected in the paired serum to analyze the potential effect of the altered microbiota community. In total, we analyzed 146 paired samples (AR = 25, PI = 43, and EF = 78). Notably, we found that the gut microbiome of AR followed a major depletion pattern with decreased 487 species and compositional diversity. Further multi-omics analysis showed depleted serum metabolites and increased inflammatory cytokines in AR and PI. Bacteroides uniformis, which declined in AR (2.4% vs 0.6%) and was negatively associated with serum IL-1ß and IL-12, was identified as a driven specie in the network of gut microbiome of EF. Functionally, the EF specimens were abundant in probiotics related to mannose and cationic antimicrobial peptide metabolism. Furthermore, a support-vector machine classifier based on microbiome, metabolome, and clinical parameters highly predicted AR (AUPRC = 0.801) and PI (AUPRC = 0.855), whereby the microbiome dataset showed a particularly high diagnostic power. In conclusion, a disruptive gut microbiota showed a significant association with allograft rejection and infection and with systemic cytokines and metabolites in LTRs.

Transcriptome analysis reveals the role of Zelda in the regulation of embryonic and wing development of Tribolium castaneum.

Zinc finger protein (Zelda) of Tribolium castaneum (TcZelda) has been showed to play pivotal roles in embryonic development and metamorphosis. However, the regulatory mechanism of TcZelda associated with these physiology processes is unclear. Herein, the developmental expression profile showed that Zelda of T. castaneum was highly expressed in early eggs. Tissue expression profiling revealed that TcZelda was mainly expressed in the larval head and adult ovary of late adults and late larvae. TcZelda knockdown led to a 95% mortality rate in adults. These results suggested that TcZelda is related to the activation of the zygote genome in early embryonic development. Furthermore, 592 differentially expressed genes were identified from the dsZelda treated group. Compared with the control group, altered disjunction (ALD) and AGAP005368-PA (GAP) in the dsZelda group were significantly down-regulated, while TGF-beta, propeptide (TGF) was significantly up-regulated, suggesting that TcZelda may be involved in insect embryonic development. In addition, the expression of Ubx ultrabithorax (UBX), Cx cephalothorax (CX), En engrailed (EN), and two Endocuticle structural glycoprotein sgabd (ABD) genes were significantly down-regulated, suggesting that they may cooperate with TcZelda to regulate the development of insect wings. Additionally, Elongation (ELO), fatty acid synthase (FAS), and fatty acyl-CoA desaturase (FAD) expression was inhibited in dsZelda insects, which could disturb the lipase signaling pathways, thus, disrupting the insect reproductive system and pheromone synthesis. These results may help reveal the function of TcZelda in insects and the role of certain genes in the gene regulatory network and provide new ideas for the prevention and control of T. castaneum.

Influence of Electrostatic Force Nonlinearity on the Sensitivity Performance of a Tapered Beam Micro-Gyroscope Based on Frequency Modulation.

Electrostatic force nonlinearity is widely present in MEMS systems, which could impact the system sensitivity performance. The Frequency modulation (FM) method is proposed as an ideal solution to solve the problem of environmental fluctuation stability. The effect of electrostatic force nonlinearity on the sensitivity performance of a class of FM micro-gyroscope is investigated. The micro-gyroscope consists of a tapered cantilever beam with a tip mass attached to the end. Considering the case of unequal width and thickness, the motion equations of the system are derived by applying Hamilton's principle. The differential quadrature method (DQM) was used to analyze the micro-gyroscope's static deflection, pull-in voltage, and natural frequency characteristics. We observed that from the onset of rotation, the natural frequencies of the drive and sense modes gradually split into a pair of natural frequencies that were far from each other. The FM method directly measures the angular velocity by tracking the frequency of the drive and sense modes. Then, based on the linear system, the reduced-order model was used to analyze the influence of the shape factor and DC voltage on the sensitivity performance. Most importantly, the nonlinear frequency of system was obtained using the invariant manifold method (IMM). The influence of electrostatic force nonlinearity on the performance of the FM micro-gyroscope was investigated. The results show that the different shape factors of width and thickness, as well as the different DC voltages along the drive and sense directions, break the symmetry of the micro-gyroscope and reduce the sensitivity of the system. The sensitivity has a non-linear trend with the rotation speed. The DC voltage is proportional to the electrostatic force nonlinearity coefficient. As the DC voltage gradually increases, the nonlinearity is enhanced, resulting in a significant decrease in the sensitivity of the micro-gyroscope. It is found that the negative shape factor (width and thickness gradually increase along the beam) can effectively restrain the influence of electrostatic force nonlinearity, and a larger dynamic detection range can be obtained.

A Mini Review of S-Nitrosoglutathione Loaded Nano/Micro-Formulation Strategies.

As a potential therapeutic agent, the clinical application of S-nitrosoglutathione (GSNO) is limited because of its instability. Therefore, different formulations have been developed to protect GSNO from degradation, delivery and the release of GSNO at a physiological concentration in the active position. Due to the high water-solubility and small molecular-size of GSNO, the biggest challenges in the encapsulation step are low encapsulation efficiency and burst release. This review summarizes the different nano/micro-formulation strategies of a GSNO related delivery system to provide references for subsequent researchers interested in GSNO encapsulation.

Cytochrome P450 gene CYP6BQ8 mediates terpinen-4-ol susceptibility in the red flour beetle, Tribolium castaneum (Herbst) (Coleoptera: Tenebrionidae).

Cytochrome P450 proteins (CYPs) in insects can encode various detoxification enzymes and catabolize heterologous substances, conferring tolerance to insecticides. This study describes the identification of a P450 gene (CYP6BQ8) from Tribolium castaneum (Herbst) and investigation of its spatiotemporal expression profile and potential role in the detoxification of terpinen-4-ol, a component of plant essential oils. The developmental expression profile showed that TcCYP6BQ8 expression was relatively higher in early- and late-larval stages of T. castaneum compared with other developmental stages. Tissue expression profiles showed that TcCYP6BQ8 was mainly expressed in the head and integument of both larvae and adults. The expression profiling of TcCYP6BQ8 in developmental stages and tissues is closely related to the detoxification of heterologous substances. TcCYP6BQ8 expression was significantly induced after exposure to terpinen-4-ol, and RNA interference against TcCYP6BQ8 increased terpinen-4-ol-induced larval mortality from 47.78 to 66.67%. This indicates that TcCYP6BQ8 may be involved in T. castaneum's metabolism of terpinen-4-ol. Correlation investigation between the CYP6BQ8 gene and terpinen-4-ol resistance in T. castaneum revealed that the TcCYP6BQ8 gene was one of the factors behind T. castaneum's resistance to terpinen-4-ol. This discovery may provide a new theoretical foundation for future regulation of T. castaneum.

LGEANet: LSTM-global temporal convolution-external attention network for respiratory motion prediction.

PURPOSE: To develop a deep learning network that treats the three-dimensional respiratory motion signals as a whole and considers the inter-dimensional correlation between signals of different directions for accurate respiratory tumor motion prediction. METHODS: We propose a deep learning framework, named as LSTM-Global Temporal Convolution-External Attention Network (LGEANet). In LGEANet, we first feed each of the univariate time series into the Long Short-Term Memory (LSTM) module respectively and utilize the strength of the global temporal convolutional layer to discover the temporal pattern of the univariate signals from hidden states of the LSTM. Then, External attention is adopted to capture the dynamic dependence of the multiple time series. Also, a traditional autoregressive linear model in parallel to the non-linear neural network part was integrated to mitigate the scale insensitivity of the networks. A total of 304 motion traces for 31 patients are acquired from a public dataset in the experiments and four representative cases were selected for model evaluation. The respiratory signals were sampled at intervals of about 37.5 ms (26 frames per second) for an average duration of 71 min. RESULTS: The proposed LGEANet achieved better performance with higher empirical correlation coefficient value (CORRs) and lower mean absolute error value (MAEs) and relative squared error value (RSEs) than other investigated models. For the four representative datasets, when the response time is less than 231 ms, the model can achieve CORRs more than 0.96. And the averaged position error reduction by using the proposed model was about 67% in the superior-inferior (SI) direction, 41% in the anterior-posterior (AP) direction and 38% in the right-left (RL) direction compared to that without prediction. The proposed network achieved the greatest error reduction in the SI direction, which is the main direction of tumor motion. CONCLUSIONS: The LGEANet achieves promising performance in minimizing the prediction error due to system latencies during real-time tumor motion tracking.

Directed-Graph-Learning-Based Diagnosis of Multiple Faults for High Speed Train With Switched Dynamics.

This article addresses the distributed multiple fault isolation, modeling, and the closed-loop fault estimation under asynchronous switching for high speed train (HST) with switched dynamics, which is composed of traction, coasting, and braking. First, directed-graph-quantum-learning-based multiple-agent system (MAS) classifiers are introduced to characterize the joints effects of multiple faults. Some sufficient conditions are derived under the condition that the multiple fault topology contains a directed spanning tree and cycle edge, and these conditions guarantee that the multiple fault isolation problem can be solved under randomized learning techniques. Then, single-integrator agents are employed to capture the time-varying topology of multiple fault modeling, in which edge agreement and persistence condition are used to guarantee asymptotic consensus. After that, a novel robust fault estimation design along with the switched Lyapunov function and average dwell time is proposed for the possible power actuator faults subject to asynchronous switching and electromagnetic interferences. In addition, switched estimators are designed such that the closed-loop system is asymptotically stable. A multiple fault isolation and estimation case is investigated to validate the application of this methodology.

Adversarial Human Trajectory Learning for Trip Recommendation.

The problem of trip recommendation has been extensively studied in recent years, by both researchers and practitioners. However, one of its key aspects-understanding human mobility-remains under-explored. Many of the proposed methods for trip modeling rely on empirical analysis of attributes associated with historical points-of-interest (POIs) and routes generated by tourists while attempting to also intertwine personal preferences-such as contextual topics, geospatial, and temporal aspects. However, the implicit transitional preferences and semantic sequential relationships among various POIs, along with the constraints implied by the starting point and destination of a particular trip, have not been fully exploited. Inspired by the recent advances in generative neural networks, in this work we propose DeepTrip-an end-to-end method for better understanding of the underlying human mobility and improved modeling of the POIs' transitional distribution in human moving patterns. DeepTrip consists of: a trip encoder (TE) to embed the contextual route into a latent variable with a recurrent neural network (RNN); and a trip decoder to reconstruct this route conditioned on an optimized latent space. Simultaneously, we define an Adversarial Net composed of a generator and critic, which generates a representation for a given query and uses a critic to distinguish the trip representation generated from TE and query representation obtained from Adversarial Net. DeepTrip enables regularizing the latent space and generalizing users' complex check-in preferences. We demonstrate, both theoretically and empirically, the effectiveness and efficiency of the proposed model, and the experimental evaluations show that DeepTrip outperforms the state-of-the-art baselines on various evaluation metrics.

MetaGeo: A General Framework for Social User Geolocation Identification With Few-Shot Learning.

Identifying the geolocation of social media users is an important problem in a wide range of applications, spanning from disease outbreaks, emergency detection, local event recommendation, to fake news localization, online marketing planning, and even crime control and prevention. Researchers have attempted to propose various models by combining different sources of information, including text, social relation, and contextual data, which indeed has achieved promising results. However, existing approaches still suffer from certain constraints, such as: 1) a very few samples are available and 2) prediction models are not easy to be generalized for users from new regions-which are challenges that motivate our study. In this article, we propose a general framework for identifying user geolocation-MetaGeo, which is a meta-learning-based approach, learning the prior distribution of the geolocation task in order to quickly adapt the prediction toward users from new locations. Different from typical meta-learning settings that only learn a new concept from few-shot samples, MetaGeo improves the geolocation prediction with conventional settings by ensembling numerous mini-tasks. In addition, MetaGeo incorporates probabilistic inference to alleviate two issues inherent in training with few samples: location uncertainty and task ambiguity. To demonstrate the effectiveness of MetaGeo, we conduct extensive experimental evaluations on three real-world datasets and compare the performance with several state-of-the-art benchmark models. The results demonstrate the superiority of MetaGeo in both the settings where the predicted locations/regions are known or have not been seen during training.