Search

The pursuit of highly efficient electrocatalysts for the alkaline hydrogen evolution reaction (HER) is of paramount importance for water splitting. However, it is still a formidable task in Mo2C-based materials because of the agglomeration and strong Mo–H binding of Mo2C units. Herein, a novel CeOCl-CeO2/Mo2C heterostructure nesting within a three-dimensional porous nitrogen-doped carbon matrix has been designed and used for catalyzing HER via simultaneous morphology and heterointerface engineering. As expected, the optimal CeOCl-CeO2(0.2)/Mo2C@3DNC exhibits impressive HER activity, with a low overpotential of 156 mV at a current density of 10 mA cm–2coupled with a slight Tafel slope of 62.20 mV dec–1. Introducing a Ce promoter, that is CeOCl and CeO2, would endow the interface with an internal electric field and electron redistribution between CeOCl-CeO2and Mo2C induced by the heterogeneous work function difference. Moreover, experimental investigation and density functional calculations confirm that the CeOCl-CeO2/Mo2C heterointerface can downshift the d-band center of the active Mo center, weakening the strength of the Mo–H coupling. This proposed concept, engineering Ce-based promoters into active entities involved in the heterostructure to modulate intermediate adsorption, offers a great opportunity for the design of superior electrocatalysts for energy conversion.

IMPORTANCE: The sustainable effectiveness and safety of a nonphysician community health care practitioner–led intensive blood pressure intervention on cardiovascular disease have not, to the authors’ knowledge, been studied, especially in the older adult population. OBJECTIVE: To evaluate such a multifaceted model with a more stringent blood pressure treatment goal (<130/80 mm Hg) among patients aged 60 years and older with hypertension. DESIGN, SETTING, AND PARTICIPANTS: This was a 48-month follow-up study of the China Rural Hypertension Control Project (CRHCP), an open-cluster randomized clinical trial, conducted from 2018 to 2023. Participants 60 years and older and younger than 60 years with a diagnosis of hypertension from the CRHCP trial were included for analysis. Individuals were recruited from 326 villages in rural China. INTERVENTIONS: The well-trained, nonphysician, community health care practitioner implemented a multifaceted intervention program (eg, initiation or titration of antihypertensive medications) to achieve a blood pressure level of less than 130/80 mm Hg, supervised by primary care physicians. MAIN OUTCOMES AND MEASURES: Cardiovascular disease (a composite of myocardial infarction, stroke, heart failure requiring hospitalization, and cardiovascular disease death). RESULTS: A total of 22 386 individuals 60 years and older with hypertension and 11 609 individuals younger than 60 years with hypertension were included in the analysis. The mean (SD) age of the participants was 63.0 (9.0) years and included 20 825 females (61.3%). Among the older individuals with hypertension, a total of 11 289 patients were randomly assigned to the intervention group and 11 097 to the usual-care group. During a median (IQR) of 4.0 (4.0-4.1) years, there was a significantly lower rate of total cardiovascular disease (1133 [2.7%] vs 1433 [3.5%] per year; hazard ratio [HR], 0.75; 95% CI, 0.69-0.81; P < .001) and all-cause mortality (1111 [2.5%] vs 1210 [2.8%] per year; HR, 0.90; 95% CI, 0.83-0.98; P = .01) in the intervention group than in the usual-care group. For patients younger than 60 years, the risk reductions were also significant for total cardiovascular disease (HR, 0.64; 95% CI, 0.56-0.75; P < .001), stroke (HR, 0.64; 95% CI, 0.55-0.76; P < .001), heart failure (HR, 0.39; 95% CI, 0.18-0.87; P = .02), and cardiovascular death (HR, 0.54; 95% CI, 0.37-0.77; P < .001), with all interaction P values for age groups greater than .05. In both age categories, the incidences of injurious falls, symptomatic hypotension, syncope, and the results for kidney outcomes did not differ significantly between groups. CONCLUSIONS AND RELEVANCE: In both the aging and younger general population with hypertension, the nonphysician health care practitioner–led, multifaceted, intensive blood pressure intervention model could effectively and safely reduce the risk of cardiovascular disease and all-cause death. TRIAL REGISTRATION: ClinicalTrials.gov Identifier: NCT03527719

It is of great interest to overcome the problem of high-temperature oxidation of press-hardened steels (PHS) in press hardening with a cost-effective Cr alloying strategy. Therefore, since Cr is one of the expensive metals, the effect of Cr (Cr ≤ 3 wt%) content on the oxidation resistance of Cr–Si–Mn alloyed steels was investigated. It was found that Cr–Si–Mn alloyed PHS with a Cr content of about 1.5 wt% or less cannot form a continuous Cr2O3layer during press hardening. However, when the Cr content reaches around 2.5 wt%, the PHS shows excellent oxidation resistance due to the formation of continuous Cr2O3and amorphous SiO2layers, as well as an Mn-rich oxide layer. However, further addition of 0.5 wt% could not significantly improve the oxidation resistance of this PHS. Moreover, the increase in Cr content simultaneously enhances the positive role of Si and Mn in high-temperature oxidation resistance. This work provides new insights into the design of the next generation of PHSs with economical and superior performance.

Vehicular edge computing (VEC), which extends the computing, storage, and networking resources from the cloud center to the logical network edge through the deployment of edge servers at the road-side unit (RSU), has aroused extensive attention in recent years, by virtue of the advantages in meeting the stringent latency requirements of vehicular applications. VEC enables the tasks and data to be processed and analyzed in close proximity to data sources (i.e., vehicles). VEC reduces the response latency for vehicular tasks, but also mitigates the burdens over the backhaul networks. However, how to achieve cost-effective task offloading in VEC remains a challenging problem, owing to the fact that the computing capabilities of the edge server are not sufficient enough compared to the cloud center and the uneven distribution of computing resources among RSUs. In this paper, we consider an urban VEC scenario and model the VEC system in terms of delay and cost. The goal of this paper is to minimize the weighted total latency and vehicle cost by balancing the bandwidth and migrating tasks while satisfying multiple constraint conditions. Specifically, we model the task offloading problem as a weighted bipartite graph matching problem and propose a Kuhn-Munkres (KM) based Task Matching Offloading scheme (KTMO) to determine the optimal offloading strategy. Furthermore, considering the dynamic time-varying features of the VEC environment, we model the task migration problem as a Markov Decision Process (MDP) and propose a Deep Reinforcement Learning (DRL) based online learning method to explore optimal migration decisions. The experimental results demonstrate that our strategy has better performance compared to other methods.

Personalized review generation is significant for e-commerce applications, such as providing explainable recommendation and assisting the composition of reviews. With the success of pre-trained language models (PLMs), prompt learning-based approaches have been employed to handle this task. However, the existing approach neglects the historical user-item interactions as well as the diverse semantics of the reviews (including semantically relevant reviews and semantically irrelevant reviews). In this paper, we propose GRAPA, a graph-enhanced prompt learning approach for personalized review generation. Specifically, GRAPA extracts topic-level information for each review to address the semantic diversity of reviews. Moreover, GRAPA employs a heterogeneous graph neural network (GNN) to explore the collaborative information hidden in historical user-item interactions. User and item representations generated by the GNN module as well as their ID embeddings are used as prompts and fed into a PLM to guide the generation process. To alleviate the interference of semantically irrelevant reviews, GRAPA further proposes a contrastive learning module to distinguish them. Experimental results on public datasets show that GRAPA outperforms existing methods by up to 4.3% in BLEU-4 and 5.4% in ROUGE2-F.

As turbine operating conditions change, the blade height and tip clearance undergo continuous alterations due to the combined effects of thermal stress, aerodynamic forces and centrifugal forces, subsequently influencing the turbine performance. To take this effect into account in turbine performance prediction, this study considers the influence of fluid-heat-structure coupling on blade height and tip clearance and establishes a one-dimensional comprehensive prediction method for multi-stage axial turbine performance considering blade height. When compared with experimental results from a four-stage axial turbine, by considering the fluid-thermal-solid coupling effects, the average relative error in total pressure ratio prediction is reduced from 3.76 % to 1.99 % and the average relative error in total temperature ratio prediction is reduced from 2.03 % to 1.26 %. Compared with the traditional flow prediction method, the prediction results of turbine characteristics considering blade height and tip clearance changes in this paper are closer to the experimental results.

The formation and preservation of the active phase of the catalysts at the triple-phase interface during CO2capture and reduction is essential for improving the conversion efficiency of CO2electroreduction toward value-added chemicals and fuels under operational conditions. Designing such ideal catalysts that can mitigate parasitic hydrogen generation and prevent active phase degradation during the CO2reduction reaction (CO2RR), however, remains a significant challenge. Herein, we developed an interfacial engineering strategy to build a new SnOxcatalyst by invoking multiscale approaches. This catalyst features a hierarchically nanoporous structure coated with an organic F-monolayer that modifies the triple-phase interface in aqueous electrolytes, substantially reducing competing hydrogen generation (less than 5%) and enhancing CO2RR selectivity (∼90%). This rationally designed triple-phase interface overcomes the issue of limited CO2solubility in aqueous electrolytes via proactive CO2capture and reduction. Concurrently, we utilized pulsed square-wave potentials to dynamically recover the active phase for the CO2RR to regulate the production of C1 products such as formate and carbon monoxide (CO). This protocol ensures profoundly enhanced CO2RR selectivity (∼90%) compared with constant potential (∼70%) applied at −0.8 V (V vs RHE). We further achieved a mechanistic understanding of the CO2capture and reduction processes under pulsed square-wave potentials via in situ Raman spectroscopy, thereby observing the potential-dependent intensity of Raman vibrational modes of the active phase and CO2RR intermediates. This work will inspire material design strategies by leveraging triple-phase interface engineering for emerging electrochemical processes, as technology moves toward electrification and decarbonization.

ERAP1 is an emerging target for a large subclass of severe autoimmune diseases known as “MHC-I-opathy”, together with tumor immunity. Nevertheless, effective inhibitors targeting ERAP1 remain a challenge. In this study, a novel food-derived natural product ERAP1-targeting inhibitor, carnosic acid, was identified, and to our knowledge, it is one of the best active compounds among the highly selective inhibitors targeting the orthosteric site of ERAP1. The results reveal that carnosic acid could bind strongly, like a key to the ERAP1 active site in the biased S1′ pocket, which is different from the binding mode of the existing orthosteric site inhibitors. HLA-B27-mediated cell modeling validated that carnosic acid has the activity to reverse the AS-associated cellular phenotype brought on by ERAP1 through inhibition. Our findings provide insights into the design of potent inhibitors against the ERAP1 orthosteric site and the discovery of a key direct target of carnosic acid.

With the increasing number of wireless devices connecting to networks and sharing spectrum resources, interference has become a major obstacle to improving network performance. Existing interference management (IM) methods treat interference as a negative factor and focus on suppressing or eliminating its impact on the transmission of intended signals. However, this often comes at the cost of consuming communication resources and degrading desired transmission performance. Therefore, the design of a cost-effective IM method that “exploits” interference is important. To achieve this goal, we propose Intelligent Reflecting Surface Assisted Interference Exploitation (IRS-IE) to realize efficient desired data transmission. IRS-IE leverages the low-cost and adaptive deployment capabilities of IRS to gather and reflect interference towards the interfered receiver (Rx). By appropriately designing the reflection coefficient of IRS, a phase shift is introduced to the incident interference, allowing the reflected interference to interact with its direct counterpart at the interfered Rx. As a result, the interfered Rx can retrieve its desired data from the mixed interference. This way, IRS-IE can make use of the interference to facilitate the desired data transmission. Our theoretical analysis and simulation results show that IRS-IE significantly improves the spectral efficiency (SE) of the interfered communication pair over the other IM methods.

This paper studies the differential flatness of time-delay systems (TDSs). Based on factorizations in the pseudo-polynomials ring, a flat output for the single-input linear commensurate TDS (namely, the delays in the system are multiplies of a certain unit delay) is constructed explicitly. Then the TDS can be converted into a high-order fully actuated system (HOFAS) model with the generalized state as the flat output, and the finite spectrum assignment (FSA) problem can be solved immediately by using the HOFAS approach, which provides some new insight into the study of the FSA problem. By parameterizing the considered TDS via the constructed flat output and applying the interpolation theory, the state trajectory planning problem and state tracking problem are solved, resulting a two-degree-of-freedom (2DOF) controller. Numerical examples demonstrates the effectiveness of the presented approach.

Nanoporous Co100-xNix(x= 1, 5, 10 and 15) solid solutions were synthesized using the chemical coprecipitation followed by hydrogen reduction method and applied as catalysts for the efficient hydrogen generation from NaBH4hydrolysis. The obtained catalysts are composed of hcp- and fcc-(Co,Ni) solid solutions phases, and exhibit the special three-dimensional nanoporous structure. With increasing the Ni content, the hydrogen generation rate (HGR) of NaBH4hydrolysis catalyzed by Co100-xNixsolid solutions decreased slightly, whereas the cycling stability of the catalyst can be improved. The HGR remains as much as 73% of the initial value after five times of catalysis for Co85Ni15. This work contributes to the development of advanced non-noble metal-based catalysts for efficient and long-term hydrogen generation from NaBH4hydrolysis.

Supplemental Digital Content is Available in the Text.

Platelet-derived growth factor receptor α (PDGFRα) is often considered as a general marker of mesenchymal cells and fibroblasts, but also shows expression in a portion of osteoprogenitor cells. Within the skeleton, Pdgfrα+mesenchymal cells have been identified in bone marrow and periosteum of long bones, where they play a crucial role in participating in fracture repair. A similar examination of Pdgfrα+cells in calvarial bone healing has not been examined. Here, we utilize Pdgfrα-CreERTM;mT/mG reporter animals to examine the contribution of Pdgfrα+mesenchymal cells to calvarial bone repair through histology and single-cell RNA sequencing (scRNA-Seq). Results showed that Pdgfrα+mesenchymal cells are present in several cell clusters by scRNA-Seq, and by histology a dramatic increase in Pdgfrα+cells populated the defect site at early timepoints to give rise to healed bone tissue overtime. Notably, diphtheria toxin-mediated ablation of Pdgfrα reporter+cells resulted in significantly impaired calvarial bone healing. Our findings suggest that Pdgfrα-expressing cells within the calvarial niche play a critical role in the process of calvarial bone repair.Graphical Abstract

Knowledge hypergraph representation learning, which projects entities and $n$n-ary relations into a low-dimensional vector space, remains a challenging area to be explored despite the ubiquity of $n$n-ary relational facts in the real world. Current methods are always extensions of those used for knowledge graphs with shallow or deep structures. However, shallow and linear models limit the extraction capacity of the latent knowledge, while deep and non-linear models lead to the overabundance of parameters. In this paper, we propose a novel knowledge hypergraph completion model called HJE, which utilizes the powerful capability of convolutional neural networks for efficient representation learning. Interaction-enhanced 3D convolution and relation-aware 2D convolution are jointly utilized by HJE to extract explicit and implicit global knowledge and semantic information effectively without compromising the translation property of the model. Moreover, HJE constructs a unified learnable embedding matrix to capture entity position information in knowledge tuples. The entity mask mechanism can naturally couple the multilinear scoring approach for $n$n-ary facts to speed up the training convergence of the model. Extensive experimental results on real datasets of knowledge hypergraphs and knowledge graphs demonstrate the superior performance of HJE compared with state-of-the-art baselines.

The shrinkage and collapse of wood cell walls during carbonization make it challenging to control the size and shape of carbonized wood (CW) through pre- or postprocessing (e.g., sawing, cutting, and milling). Herein, a shape-adaptive MXene shell (MS) is created on the surface of the wood cell walls. The MS limits the deformation of wood cell walls by spatial confinement and traction effects, which is supported by the inherent dimensional stability of the MS and the formation of new C–O–Ti covalent bonds between the wood cell wall and MS. Consequently, the volumetric shrinkage ratio of CW encapsulated by the MS (CW-MS) is significantly reduced from 54.8% for CW to 2.6% for CW-MS even at 800 °C. The harnessing of this collapse enables the production of CW-MS with prolonged stability and high electric conductivity (384 S m–1). These properties make CW-MS suitable for energy storage devices with various designed shapes, matching the increasingly compact and complex structures of electronic devices.

Detection of early neoplastic lesions is crucial for improving the survival rates of patients with gastric cancer. Optical enhancement mode 2 is a new image‐enhanced endoscopic technique that offers bright images and can improve the visibility of neoplastic lesions. This study aimed to compare the detection of neoplastic lesions with optical enhancement mode 2 and white‐light imaging (WLI) in a high‐risk population. In this prospective multicenter randomized controlled trial, patients were randomly assigned to optical enhancement mode 2 or WLI groups. Detection of suspicious neoplastic lesions during the examinations was recorded, and pathological diagnoses served as the gold standard. A total of 1211 and 1219 individuals were included in the optical enhancement mode 2 and WLI groups, respectively. The detection rate of neoplastic lesions was significantly higher in the optical enhancement mode 2 group (5.1% vs. 1.9%; risk ratio, 2.656 [95% confidence interval, 1.630–4.330]; p< 0.001). The detection rate of neoplastic lesions with an atrophic gastritis background was significantly higher in the optical enhancement mode 2 group (8.6% vs. 2.6%, p< 0.001). The optical enhancement mode 2 group also had a higher detection rate among endoscopists with different experiences. Optical enhancement mode 2 was more effective than WLI for detecting neoplastic lesions in the stomach, and can serve as a new method for screening early gastric cancer in clinical practice. United States National Library of Medicine (https://www.clinicaltrials.gov), ID: NCT040720521.

Terrestrial water storage (TWS) variations are associated with water mass movements, which may cause the deformation displacements of the Global Navigation Satellite System (GNSS) stations. This study investigates the spatio-temporal TWS variations and addresses the relationship between deformation variations observed in the Huang-Huai-Hai River Basin (HHHRB) and local hydrological features. Results indicate that the vertical velocities at the GNSS stations induced by TWS changes are relatively small, and the impacts of the terrestrial water storage changes are mainly reflected in the changes of seasonal characteristics. Although there is a downward TWS trend from 2011 to 2022 in most HHHRB areas, velocities from the vertical displacements of both Gravity Recovery and Climate Experiment (GRACE) and GRACE Follow-On (GFO) and the GNSS reflect that the HHHRB is undergoing an uplift process, while the magnitude of the GRACE/GFO derived velocities is much smaller than that of the GNSS solutions. Common hydrological deformations estimated from GRACE/GFO and GNSS measurements reveal that the TWS-derived displacements can explain 54.5% of the GNSS seasonal variations, with the phases of terrestrial water storage advancing by about one month relative to GNSS common signal phases. Moreover, the decrease of the groundwater storage in the HHHRB has been accelerating since 2008. After reaching its lowest level around mid-2020, it began to rise rapidly, which might be closely related to the implementation of the South-North Water Transfer Central Project.

Strain sensors which are embeded in AMOLED are introduced and studied in this paper. High accuray strain sensing are achieved, resolution of the sensors are about 0.05%. By combing the strain sensors in the panel, mechanical behaviors can be monitored and studied in real time. The snesors are used to conduct the foldable AMOLED mechanical and reliability study. Dynamic and static strains and related mechanism are introduced and discussed. Finite element simulation models considering the mechanical properties of materials including hyperelasticity, viscoelasticity and creep are established. Finally, the strain sensors and finite element anlysis are combined to study the mechanical response of AMOLED under dynamic folding, static folding, and different temperature and humidity environments.

By considering rollable AMOLED display and application, optical and mechanical performances are discussed and optimized solutions are introduced. Different COE designs for rollable AMOLED are compared in power consumption, L‐decay and reflection. A rollable AMOLED module structure is designed with the impact absorption layer of the hybrid cover. Based on finite element simulation and neutral layer theory, module stack is optimized, accumuled strain is eliminated. Rollable performance and impact resistance are balanced. A portable device with rollable AMOLED is developed. Its mechanical design for rollable display application is introduced.

We have proposed the “Standard 4‐edge Curved Model” for the first time, which contains a standard geometric model based on superelliptic curves and a mapping relationship from a plane to a three‐dimensional surface. In this model, we can quantitatively compare the wrinkling risks of almost all known 4‐edge curved displays and provide the optimizations for the geometric parameters.

Objective: Abnormal expression of T-lymphokine-activated killer cell-originated protein kinase (TOPK) was reported to be closely related to the resistance of prostate cancer to radiotherapy and to targeted drug resistance in lung cancer. However, the role of TOPK inhibition in enhancing radiosensitivity of colorectal cancer (CRC) cells is unclear. This study aimed to evaluate the radiosensitization of TOPK knockdown in CRC cells. Methods: The expression of TOPK was detected in CRC tissues by immunohistochemistry, and the effect of TOPK knockdown was detected in CRC cells by Western blotting. CCK-8 and clonogenic assays were used to detect the growth and clonogenic ability of CRC cells after TOPK knockdown combined with radiotherapy in CRC cells. Furthermore, proteomic analysis showed that the phosphorylation of TOPK downstream proteins changed after radiotherapy. DNA damage was detected by the comet assay. Changes in the DNA damage response signaling pathway were analyzed by Western blotting, and apoptosis was detected by flow cytometry. Results: The expression of TOPK was significantly greater in CRC tissues at grades 2–4 than in those at grade 1. After irradiation, CRC cells with genetically silenced TOPK had shorter comet tails and reduced expression levels of DNA damage response-associated proteins, including phospho-cyclin-dependent kinase 1 (p-CDK1), phospho-ataxia telangiectasia-mutated (p-ATM), poly ADP-ribose polymerase (PARP), and meiotic recombination 11 homolog 1 (MRE11). Conclusions: TOPK was overexpressed in patients with moderately to poorly differentiated CRC. Moreover, TOPK knockdown significantly enhanced the radiosensitivity of CRC cells by reducing the DNA damage response.

Forest age structure, shaped by past land-use and land-cover changes (LULCC), is pivotal for estimating ecosystem carbon sinks. China’s extensive LULCC in recent decades has led to a complex forest age structure, but its impact on the carbon sink remains uncertain. Here, using a process-based ecosystem model with an explicit representation of forest age cohorts, we estimate China’s terrestrial carbon sink as 198 ± 54 TgC yr−1in the 2010s, mainly contributed by middle-aged forests. The existing forests in 2020 contribute most to the future total carbon sink, but its contribution will decrease significantly by −1.1∼−0.35 TgC yr−1until 2100 due to forest aging and the slowdown of CO2concentration growth. Future re/afforestation will enhance carbon sink by increasing forest area and rejuvenating forest demography. Our study emphasizes the limited future carbon sink due to forest aging, implying that realizing China’s carbon neutrality should not rely excessively on ecosystem carbon sink.

With the increasing demands for crop yield and production mechanization, improvement of plant architecture is getting more imperative in cotton. In the present study, we investigated the plant height (PH) and the node of first fruiting branch (NFFB) of 719 upland cotton accessions in six different environments. We used the 10,511 high‐quality single nucleotide polymorphisms (SNPs) to perform single‐locus and multi‐locus genome‐wide association study (GWAS). As many as 278 associated SNPs were identified, 142 by the mixed linear model (MLM) in the single‐locus model and 192 by six multi‐locus models. A total of 42 SNPs were identified in at least four environment‐traits or methods, of which 28 SNPs were significantly associated with PH, 22 were significantly associated with NFFB, and eight were co‐associated with the two traits. Notably, most of loci were novel besides i33922Gh located in the reported QTL for PH. Furthermore, we identified nine promising candidate genes, among which the three genes Gh_D03G0738, Gh_D10G2028, and Gh_D05G3600contained non‐synonymous SNP mutation. The accessions with alleles of the mutations resulted in significant phenotypic differences. The expression of these genes showed significant differences between short‐PH and high‐PH varieties. Moreover, overexpression of Gh_D03G0738led to reduction of PH in Arabidopsis. These results provided insights into genetic basis of plant architecture in cotton. We identified 278 single nucleotide polymorphisms (SNPs) associated plant height (PH) and node of the first fruiting branch with single‐ and multi‐locus genome‐wide association study by using 719 cotton accessions.A total of 42 SNPs were identified in at least four environment‐traits or methods.Overexpression of Gh_D03G0738led to reduction of PH in Arabidopsis.

Cysteine (Cys), as one of the biological thiols, is related to many physiological and pathological processes in humans and plants. Therefore, it is necessary to develop a sensitive and selective method for the detection and imaging of Cys in biological organisms. In this work, a novel near-infrared (NIR) fluorescent probe, Probe-Cys, was designed by connecting furancarbonyl, as a new recognition moiety, with Fluorophore–OHvia the decomposition of IR-806. The use of the furan moiety is anticipated to produce more effective fluorescence quenching because of the electron-donating ability of the O atom. Probe-Cyshas outstanding properties, such as a new recognition group, an emission wavelength in the infrared region at 710 nm, a linear range (0–100 μM), a low detection limit of 0.035 μM, good water solubility, excellent sensitivity, and selectivity without the interference of Hcy, GSH, and HS–. More importantly, Probe-Cyscould achieve the detection of endogenous Cys by reacting with the stimulant 1,4-dimercaptothreitol (DTT) and the inhibitor N-ethylmaleimide (NEM) in HepG2 cells and zebrafish. Ultimately, it was successfully applied to obtain images of Arabidopsis thaliana, revealing that the content of Cys in the meristematic zone was higher than that in the elongation zone, which was the first time that the NIR fluorescence probe was used to obtain images of Cys in A. thaliana. The superior properties of the probe exhibit its great potential for use in biosystems to explore the physiological and pathological processes associated with Cys.

This article reported the application of wrist-ankle acupuncture with the ultrasound-guided interfascial plane block (Transversus abdominis plane block,TAPB; Rectus sheath block, RSB; Rhomboid intercostal and sub–serratus block, RISSB) in two emergency laparotomy in elderly frail patients. These cases were successfully completed with only a small dosage of narcotic analgesics, and the patient's spontaneous breathing was preserved and the vital signs of the patients were stable with the Visual Analogue Scale (VAS) of 2 points and the oxygen saturation level over 95%. Both patients recovered soon after the surgery. The combination of wrist-ankle acupuncture and interfascial plane block to preserve spontaneous breathing during laparotomy in elderly frail patients in China had not been previously reported.

Terahertz (THz) waves have been widely hailed as a key enabling technology for future sixth generation (6G) wireless networks. Dynamic modulation of their polarization states is of great attraction for high-capacity communications and anisotropic sensing. The development of such technology is, however, still in very early stage owing to the difficulties of realizing electrical reconfigurability for THz devices. Artificially constructed metasurfaces and new nanomaterials, such as graphene, have been shown to provide revolutionary platforms for manipulating and controlling the wave properties, especially at THz frequencies. This work leverages the light–matter interaction in a graphene-integrated metasurface functioning as an electrically reconfigurable THz polarization converter. A novel graphene-gold bilayer topology is applied to construct such a metasurface which enables wide-range electrical tunability of the polarization conversion. Under a y-polarized illumination, the reflected components of x- and y-polarizations are tuned dynamically through an external bias voltage across the metasurface, thereby producing an elliptically polarized wave with tuneable ellipticity and angle. By changing the voltage from 0 V to 12 V, the reflected polarization ellipticity has been tuned from −0.94 to −0.5 at around 240 GHz, featuring linear-to-circular and linear-to-elliptical polarization conversions. Meanwhile, the polarization angle has been modulated from 12° to −23° at around 236 GHz. This work provides an experimentally validated THz graphene-integrated metasurface with wide polarization modulation depths, low biasing voltages and simple configuration. It promises great potential for applications in future THz communications and sensing.

The technical problem of the difficult wetting of coal bed water injection urgently needs to be solved, and the wetting of surfactants on the coal body is mainly affected by the concentration. In recent years, research related to surface-active ionic liquids (SAILs) with long-chain alkyl groups has attracted much attention. Salt can enhance the wetting effect of surfactants, but there are few studies on the effect of inorganic salts on the surface activity of [C12MIm]Cl and the effect of compound wetting bituminous coal. In this paper, NaCl was selected to be complexed with [C12MIm]Cl, and the trends of key functional groups and surface activity changes during coal wetting were quantitatively characterized from a microscopic point of view by functional group and wetting characterization analysis. The results showed that under the treatment of 0.6 mol/L of NaCl + 10 mmol/L of the [C12MIm]Cl compound, the surface tension was reduced by 7 N m/m and the minimum contact angle by 15.5°. In the modified coal, the peak area of the ═CH2group is reduced by 29%, the peak area of the −CH3group is increased by 24%, the peak area of the OH···N group is reduced by 14%, and the wetting effect is enhanced. However, the high concentration of NaCl in the composite solution will weaken the wetting effect on bituminous coal. The research results provide theoretical support for coal seam composite wetting by water injection.

Chronic pancreatitis (CP) is a multifaceted disorder influenced by environmental and genetic factors, with smoking and alcohol consumption being major contributors. Recent developments encompass the advent of innovative transgenic models and the identification of susceptibility genes, shedding light on the genetic aspect of CP. The pathogenesis of this disease involves a complex interplay of pancreatic acinar cell dysfunction, inflammatory reactions, and fibrosis. Current research delves into understanding these molecular mechanisms. Pain, a pivotal symptom of CP, has been increasingly studied to develop effective therapeutic interventions. Diagnostic advancements, including endoscopic ultrasound, radiomics, and blood-based markers, have shown potential in enhancing early CP detection. Moreover, recent clinical trials have optimized treatment approaches, such as pancreatic stone fragmentation, stent placement, and decision-making between endoscopic and surgical procedures. Emerging therapies, including chemical pancreatectomy and gene therapy, present promising opportunities for improved CP management.

This paper proposes an optimal, robust, and efficient guidance scheme for the perturbed minimum-time low-thrust transfer toward the geostationary orbit. The Earth's oblateness perturbation and shadow are taken into account. It is difficult for a Lyapunov-based or trajectory-tracking guidance method to possess multiple characteristics at the same time, including high guidance optimality, robustness, and onboard computational efficiency. In this work, a concise relationship between the minimum-time transfer problem with orbital averaging and its optimal solution is identified, which reveals that the five averaged initial costates that dominate the optimal thrust direction can be approximately determined by only four initial modified equinoctial orbit elements after a coordinate transformation. Based on this relationship, the optimal averaged trajectories constituting the training dataset are randomly generated around a nominal averaged trajectory. Five polynomial regression models are trained on the training dataset and are regarded as the costate estimators. In the transfer, the spacecraft can obtain the real-time approximate optimal thrust direction by combining the costate estimations provided by the estimators with the current state at any time. Moreover, all these computations onboard are analytical. The simulation results show that the proposed guidance scheme possesses extremely high guidance optimality, robustness, and onboard computational efficiency.

Multiple pieces of evidence have shown that prenatal exposure to perfluoroalkyl and polyfluoroalkyl substances (PFASs) is closely related to adverse birth outcomes for infants. However, difficult access to human samples limits our understanding of PFASs transport and metabolism across the human placental barrier, as well as the accurate assessment of fetal PFASs exposure. Herein, we assess fetal exposure to 28 PFASs based on paired serum, placenta, and meconium samples. Overall, 21 PFASs were identified first to be exposed to the fetus prenatally and to be metabolized and excreted by the fetus. In meconium samples, 25 PFASs were detected, with perfluorooctane sulfonate and perfluorohexane sulfonic acid being the dominant congeners, suggesting the metabolism and excretion of PFASs through meconium. Perfluoroalkyl sulfonic acids might be more easily eliminated through the meconium than perfluorinated carboxylic acids. Importantly, based on molecular docking, MRP1, OATP2B1, ASCT1, and P-gp were identified as crucial transporters in the dynamic placental transfer of PFASs between the mother and the fetus. ATSC5p and PubchemFP679 were recognized as critical structural features that affect the metabolism and secretion of PFASs through meconium. With increasing carbon chain length, both the transplacental transfer efficiency and meconium excretion efficiency of PFASs showed a structure-dependent manner. This study reports, for the first time, that meconium, which is a noninvasive and stable biological matrix, can be strong evidence of prenatal PFASs exposure.

Mechanism of cerium-based inclusion formation and influence on impact toughness were researched in industrial Al-killed steel. In the final product, actual inclusions were complex, consisting of Al-O-Ce (Ce-poor), Al-O-Ce (Ce-rich), Al-O-Ce-S(Al-poor) and S-Ca (Mn-rich) from inside to outside. Results of thermodynamic balance calculation were completely inconsistent with the actual inclusions. Considering decomposing or disappearing of formed inclusions in steel difficult, new computing results showed that inclusions were formed by firstly Al2O3(exist in molten steel before adding Ce), AlCeO3, Ce2O2S, CaS, Ca (Mn)S and MnS in turn, in better agreement with the actual inclusions. Outside of the actual inclusions, Ce2O2S and CaS had a smaller disregistry with ferrite, which could offer effective nucleation sites. And small inclusion was surrounded in a grain and bigger inclusion was surrounded in two grains at least. Microstructures were observed that grains with Ce were finer than grains without Ce. Moreover, V-notch impact toughness of samples was tested at 0 and −60°C, respectively. Compared with samples without Ce, the values of impact toughness were increased by 161% at 0°C and 225% at −60°C, respectively. And fracture surface with Ce had more and finer dimples and finer cleavages than that without Ce.

As a genetics-based machine learning technique, zeroth-level classifier system (ZCS) based on average reward reinforcement learning (ZCSAR) evolves solutions to optimize average reward per time step. However, initial experimental results have shown that, in some cases, the performance of ZCSAR oscillates heavily during the learning period, or cannot reach the optimum during the testing period. In this paper, we modify the selection strategies in ZCSAR to improve its performance, under conditions of minimal changes of ZCSAR. The proposed selection strategies take tournament selection method to choose parents in Genetic Algorithm (GA), and take roulette wheel selection method to choose actions in match set and to choose classifiers for deletion in both GA and covering. Experimental results show that ZCSAR with the new selection strategies can evolve more promising solutions with enough parameter independence, and also with slighter oscillation during the learning period.

In recent years, the full-touch human–machine interface (HMI) mode has been widely used in vehicles built by Tesla. This interaction mode replaces the conventional physical interaction modality with a screen, and it has a good sense of technological experience. However, it is unclear whether this mode will make the driver's lateral control more challenging than the conventional mode (CM). To investigate this issue, two most common secondary tasks were designed: dialing and navigation entry tasks and real-world road experiments were conducted using two instrumented vehicles. The vehicle operating parameters and the driver manual data were collected in different modes, respectively. Interestingly, the opposite results were found regarding the effect of the full-touch mode (FTM) on the driver's lateral control ability in different secondary tasks. Compared with the CM, the lateral control ability decreased less during the dialing task relative to the baseline driving in the FTM, while the lateral control ability decreased more in the FTM during the navigation entry task. In addition, drivers’ lateral control decreased further as task difficulty and driving speed increased regardless of mode. This study provides a theoretical basis for the development of laws and regulations regarding full-touch HMI mode.

Cancer-cell-specific fluorescent photosensitizers (PSs) are highly desired molecular tools for cancer ablation with minimal damage to normal cells. However, such PSs that can achieve cancer specification and ablation and a self-reporting manner concurrently are rarely reported and still an extremely challenging task. Herein, we have proposed a feasible strategy and conceived a series of fluorescent PSs based on simple chemical structures for identifying and killing cancer cells as well as monitoring the photodynamic therapy (PDT) process by visualizing the change of subcellular localization. All of the constructed cationic molecules could stain mitochondria in cancer cells, identify cancer cells specifically, and monitor cancer cell viability. Among these, IVP-Br has the strongest ability to produce ROS, which serves as a potent PS for specific recognition and killing of cancer cells. IVP-Br could translocate from mitochondria to the nucleolus during PDT, self-reporting the entire therapeutic process. Mechanism study confirms that IVP-Br with light irradiation causes cancer cell ablation via inducing cell cycle arrest, cell apoptosis, and autophagy. The efficient ablation of tumor through PDT induced by IVP-Br has been confirmed in the 3D tumor spheroid chip. Particularly, IVP-Br could discriminate cancer cells from white blood cells (WBCs), exhibiting great potential to identify circulating tumor cells (CTCs).

Background: Energy consumption is a common problem in hospital buildings, which consume twice that of other public buildings. Therefore, it is of great significance to study the control strategy for the efficient operation of the cold source system.Objective: The study aimed to explore an efficient operation control strategy for cold source system, and new technologies and patents have emerged for the same. This work, utilizing EnergyPlus, modelled and analyzed the cold source system in a Shanghai hospital to optimize its operation.Methods: The accuracy of the simulation was verified by comparing it with experimental data. Based on the simulation results, the factors influencing the energy efficiency of the cold source system were analyzed, and then the operation control strategy of the cold source system was obtained. The XGBoost was used to fit the simulation results, and the operation strategy under full operating conditions was obtained.Results: The simulated results indicated the average energy saving rates during the summer season of the chillers, the chilled water pumps, the cooling water pumps, and the cooling towers to be 6.5%, -4.0%, 38.3%, and 5.4%, respectively, under the optimal operation control strategy. The average system Coefficient of Performance (COP) of the cold source system was 5.9, and the total energy consumption was 957016.3 kW·h, which was 7.1 % energy saving compared to that under the original operation.Conclusion: The conclusions of this study could provide references for the hospital buildings’ cold source system and group control method. This study has important practical significance for the efficient control strategy of cold source systems.

The process states in gas–liquid flow are the synthetical consequence of multiple apparent factors related to time, space, flow conditions, and so on. Higher order tensor decomposition in multilinear algebra provides a powerful framework mathematically to analyze the multifactor structure of flow state ensembles. Focusing on entangling the constitute factors or modes embedded in various states, a novel strategy about graph-embedding-based multilinear subspace analysis (MSA) of state ensembles in gas–liquid flow is proposed. Pulse-wave ultrasonic Doppler (PWUD) sensor is adopted to acquire spatiotemporal information of flow states under varying flow conditions. Under the architecture of graph embedding, tensor dynamic local preservation projection (TDLPP) is constructed via PWUD signal to further describe the latent characteristics of various flow states in time–space dimensions. In the reconstructed graph embedding latent space, higher order singular value decomposition (HOSVD) is used for tensor analysis of flow state ensembles that combine several modes, including different state categories, time, space, and flow conditions. The purpose of the strategy is to further achieve state monitoring of typical and transitional processes in gas–liquid flow under varying flow conditions with a simple sensor, on the foundation of refining apparent factors underlying flow state formation. Dynamic experiments are designed to verify the effectiveness of the proposed scheme.

Background: Hospitals account for the most proportion of energy consumption in the public building sector. Chillers usually account for most of the overall energy consumption of the cold source system.Objective: To solve the problem of chillers' large energy consumption problem, novel technologies were developed, and achievements were patented.Methods: The operating characteristics influencing factors of the magnetic suspension centrifugal chiller (MSCC) and variable frequency screw chiller (VFSC) of a hospital in Shanghai were analyzed and discussed by actual measurements. Then, based on the operating characteristics of the chiller obtained from the analysis of the measured data, the cooling capacity was classified by the K-Means clustering method to obtain the startup strategy of the chillers.Results: The effects of the supply chilled water temperature, the supply cooling water temperature and variable cooling water flow rate on the maximum cooling capacity and coefficient of performance (COP) of both chillers were explored. The load distribution scheme was discussed based on the chillers' startup strategy.Conclusion: The average part load ratio operation scheme was the preferred chiller distribution scheme. A chiller's maximum allowable part load ratio should not exceed 80% during the low-load operation period and should not be less than 60% during the conventional operation period. It provided a reference for optimizing the chiller operation strategy to reduce system energy consumption.

Recently, a new research trend in our video salient object detection (VSOD) research community has focused on enhancing the detection results via model self-fine-tuning using sparsely mined high-quality keyframes from the given sequence. Although such a learning scheme is generally effective, it has a critical limitation, i.e., the model learned on sparse frames only possesses weak generalization ability. This situation could become worse on “long” videos since they tend to have intensive scene variations. Moreover, in such videos, the keyframe information from a longer time span is less relevant to the previous, which could also cause learning conflict and deteriorate the model performance. Thus, the learning scheme is usually incapable of handling complex pattern modeling. To solve this problem, we propose a divide-and-conquer framework, which can convert a complex problem domain into multiple simple ones. First, we devise a novel background consistency analysis (BCA) which effectively divides the mined frames into disjoint groups. Then for each group, we assign an individual deep model on it to capture its key attribute during the fine-tuning phase. During the testing phase, we design a model-matching strategy, which could dynamically select the best-matched model from those fine-tuned ones to handle the given testing frame. Comprehensive experiments show that our method can adapt severe background appearance variation coupling with object movement and obtain robust saliency detection compared with the previous scheme and the state-of-the-art methods.

In this work, the influence of high- ${k}$ /metal gate (HKMG) thermal processes such as post dielectric annealing (PDA), post metal annealing (PMA), and post amorphous Si cap annealing (PCA) on metal boundary effect (MBE) in FinFET is investigated. It is revealed that the PDA temperature increase leads to more severe MBE. On the contrary, the increase of soak or spike temperature of PMA is beneficial to reduce MBE. A higher PCA spike temperature also reduces MBE. It is demonstrated that by using an optimized combination of these three anneals, the threshold voltage ( ${V}_{\text {t}}{)}$ shift induced by MBE can be reduced by about 50% without sacrificing device performance and reliability.

Grouting has been widely adopted in disaster management related to tunneling. Since completely and strongly weathered granite is problematic stratum in tunneling construction, grouting simulation tests were carried out to investigate the grout diffusion pattern with various particle size distributions and grouting pressures. Injected grout mass and characteristics of the consolidating body were obtained to reveal the mechanism of the grouting process. Permeation, compaction, fracture, and their mixed diffusion pattern were defined, and the different diffusion patterns were distinguished by the particle size distribution and grouting pressure. The compaction pattern and the permeation pattern have the lowest and highest injected grout mass, respectively. Injected grout mass increased with the increment of grouting pressure. Particle size distribution and grouting pressure determine the grout diffusion pattern. The permeation-dominated pattern can be distinguished from compaction and fracture pattern by D10, while the compaction and fracture patterns are differentiated by D60. The grout diffusion pattern is determined by grouting pressure in the specific strata though it has no relationship with the groutability. Permeation pattern has the highest reinforcement effect with a “sandcrete” structure, the compaction pattern has the lowest reinforcement effect with compaction effect, and the fracture pattern enhances strata by compaction and skeleton support. Finally, a criterion was obtained to forecast the completely and strongly weathered granite’s grout diffusion pattern, which was confirmed by the grouting engineering.

Recently, ChatGPT, a representative large language model (LLM), has gained considerable attention. Due to their powerful emergent abilities, recent LLMs are considered as a possible alternative to structured knowledge bases like knowledge graphs (KGs). However, while LLMs are proficient at learning probabilistic language patterns and engaging in conversations with humans, they, like previous smaller pre-trained language models (PLMs), still have difficulty in recalling facts while generating knowledge-grounded contents. To overcome these limitations, researchers have proposed enhancing data-driven PLMs with knowledge-based KGs to incorporate explicit factual knowledge into PLMs, thus improving their performance in generating texts requiring factual knowledge and providing more informed responses to user queries. This paper reviews the studies on enhancing PLMs with KGs, detailing existing knowledge graph enhanced pre-trained language models (KGPLMs) as well as their applications. Inspired by existing studies on KGPLM, this paper proposes enhancing LLMs with KGs by developing knowledge graph-enhanced large language models (KGLLMs). KGLLM provides a solution to enhance LLMs’ factual reasoning ability, opening up new avenues for LLM research.

Temporally incomparability across the scan lines in polar-orbiting satellite-derived land surface temperature (LST) affects their widespread application. Some challenges persist in the existing research on this issue, such as the absence of a universal algorithm applicable for the partly clear-sky condition in the daytime and scale inconsistency of the used datasets, when LST varies nonlinearly over time. Given this situation, we proposed an improved approach for temporal normalization, integrating ensemble regression models and a new LST variation rate model (RM), which captures typical LST variation characteristics over time during polar-orbiting satellite overpass periods. The Aqua Moderate Resolution Imaging Spectroradiometer (MODIS) LST data across the contiguous United States (CONUS) were collected to investigate its effectiveness. Moreover, cross-validation was conducted using the time-interpolated Geostationary Operational Environmental Satellite 16 (GOES-16) advanced baseline imager (ABI) LST. The normalized LST had remarkable consistency with the GOES-16 LST, with superior accuracy in contrast with the original LST. The root-mean-squared error (RMSE) was improved by approximately 1.56 K, and bias was enhanced up to 1.80 K. This study exhibited relatively superior performance in terms of quantitative outcomes and spatial distribution of LST compared with the previous studies. These evaluations indicate that the proposed method could be a dependable and general solution for addressing temporal inconsistencies in clear-sky LST during polar-orbiting satellite overpass periods.

Lower limb motion recognition techniques commonly employ Surface Electromyographic Signal (sEMG) as input and apply a machine learning classifier or Back Propagation Neural Network (BPNN) for classification. However, this artificial feature engineering technique is not generalizable to similar tasks and is heavily reliant on the researcher’s subject expertise. In contrast, neural networks such as Convolutional Neural Network (CNN) and Long Short-term Memory Neural Network (LSTM) can automatically extract features, providing a more generalized and adaptable approach to lower limb motion recognition. Although this approach overcomes the limitations of human feature engineering, it may ignore the potential correlation among the sEMG channels. This paper proposes a spatial–temporal graph neural network model, STGNN-LMR, designed to address the problem of recognizing lower limb motion from multi-channel sEMG. STGNN-LMR transforms multi-channel sEMG into a graph structure and uses graph learning to model spatial–temporal features. An 8-channel sEMG dataset is constructed for the experimental stage, and the results show that the STGNN-LMR model achieves a recognition accuracy of 99.71%. Moreover, this paper simulates two unexpected scenarios, including sEMG sensors affected by sweat noise and sudden failure, and evaluates the testing results using hypothesis testing. According to the experimental results, the STGNN-LMR model exhibits a significant advantage over the control models in noise scenarios and failure scenarios. These experimental results confirm the effectiveness of the STGNN-LMR model for addressing the challenges associated with sEMG-based lower limb motion recognition in practical scenarios.

Pneumonia is one of the largest causes of death in the world. Deep learning techniques can assist doctors to detect the areas of pneumonia in the chest X-rays images. However, existing methods lack sufficient consideration for the large variation scale and the blurred boundary of the pneumonia area. Here, we present a deep learning method based on Retinanet for pneumonia detection. First, we introduce Res2Net into Retinanet to get the multi-scale feature of pneumonia. Then, we proposed a novel predicted boxes fusion algorithm, named Fuzzy Non-Maximum Suppression (FNMS), which gets a more robust predicted box by fusing the overlapping detection boxes. Finally, we get the performance outperforms than existing methods by integrating two models with different backbones. We report the experimental result in the single model case and the model ensemble case. In the single model case, RetinaNet with FNMS algorithm and Res2Net backbone is better than RetinaNet and other models. In the model ensemble case, the final score of predicted boxes that fused by the FNMS algorithm is better than NMS, Soft-NMS, and weighted boxes fusion. Experimental results on the pneumonia detection dataset verify the superiority of the FNMS algorithm and the proposed method in the pneumonia detection task.

Objective: Autosomal recessive bestrophinopathy (ARB), a retinal degenerative disease, is characterized by central visual loss, yellowish multifocal diffuse subretinal deposits, and a dramatic decrease in the light peak on electrooculogram. The potential pathogenic mechanism involves mutations in the BEST1gene, which encodes Ca2+-activated Cl−channels in the retinal pigment epithelium (RPE), resulting in degeneration of RPE and photoreceptor. In this study, the complete clinical characteristics of two Chinese ARB families were summarized. Methods: Pacific Biosciences (PacBio) single-molecule real-time (SMRT) sequencing was performed on the probands to screen for disease-causing gene mutations, and Sanger sequencing was applied to validate variants in the patients and their family members. Results: Two novel mutations, c.202T>C (chr11:61722628, p.Y68H) and c.867+97G>A, in the BEST1gene were identified in the two Chinese ARB families. The novel missense mutation BEST1c.202T>C (p.Y68H) resulted in the substitution of tyrosine with histidine in the N-terminal region of transmembrane domain 2 of bestrophin-1. Another novel variant, BEST1c.867+97G>A (chr11:61725867), located in intron 7, might be considered a regulatory variant that changes allele-specific binding affinity based on motifs of important transcriptional regulators. Conclusion: Our findings represent the first use of third-generation sequencing (TGS) to identify novel BEST1mutations in patients with ARB, indicating that TGS can be a more accurate and efficient tool for identifying mutations in specific genes. The novel variants identified further broaden the mutation spectrum of BEST1in the Chinese population.

Background: Endoscopic ultrasound-guided pancreatic duct (PD) drainage (EUS-PDD) is being increasingly performed as an alternative method to surgical drainage to achieve PD decompression after failed endoscopic retrograde pancreatography (ERP). However, no directly study has compared EUS-PDD with surgical PD drainage after failed ERP in patients with chronic pancreatitis. Methods: Consecutive patients who underwent EUS-PDD or longitudinal pancreaticojejunostomy after failed ERP were retrospectively identified from our endoscopy and medical information systems. The primary end point was the Izbicki pain score. The secondary end points were pain relief at the end of follow-up, procedure outcomes, adverse events, readmission, and reintervention. Results: A total of 21 patients (11 EUS-PDD, 10 surgical drainages) were analyzed. There were no significant differences in mean Izbicki pain score (EUS-PDD, 13.6 ± 10.1 vs. surgical drainage 10.7 ± 7.9, p= 0.483) or complete/partial pain relief (60%/30% vs. 70%/30%, p= 0.752) at the end of follow-up of the two groups. The rates of overall adverse events (27.3% vs. 30.0%, p= 0.893) and readmission (63.6% vs. 40.0%, p= 0.290) were similar in the two treatment groups, while patients in EUS-PDD group required more reinterventions (45.5% vs. 0%, p= 0.039) compared with patients in the surgery group. Conclusion: EUS-PDD showed comparable pain relief and safety to surgical PD drainage after failed ERP, with a higher rate of reintervention. The selection of EUS-PDD or surgical drainage may be appropriate based on an individualized strategy. Graphical abstract:

Next destination recommendation is a crucial research area for understanding human travel behavior. However, existing studies often overlook the problem of underfitting, which arises due to the limited regularity in users’ travel patterns. To tackle this issue, we leverage diverse co-occurrence patterns (CoPs) to discover potential user preferences. These patterns capture intersections with similar spatial and temporal characteristics in users’ travels. However, traditional graph neural network (GNN)-based approaches struggle to effectively handle complex spatial-temporal CoPs. To overcome these challenges, we propose a novel framework called DHIN (Dynamic Heterogeneous Information Network). First, to address the problem of underfitting, DHIN constructs intricate CoPs by leveraging abundant features and connection relationships. Additionally, to solve the needs of cold-start users, DHIN generates potential connections by capturing dynamic urban hotspots based on global users’ travel trajectories. Moreover, to model dynamic heterogeneous information, DHIN utilizes a hierarchical attention mechanism and integrates a dynamic encoder. The mechanism integrates multi-level attention to learn informative embeddings from heterogeneous attributes and structures, while the dynamic encoder processes dynamic temporal information for updating node representations. Finally, extensive experiments conducted on real-world trajectory data demonstrate the effectiveness of the proposed DHIN model.