Search

Previous works employ the Large Language Model (LLM) like GPT-3 for knowledge-based Visual Question Answering (VQA). We argue that the inferential capacity of LLM can be enhanced through knowledge injection. Although methods that utilize knowledge graphs to enhance LLM have been explored in various tasks, they may have some limitations, such as the possibility of not being able to retrieve the required knowledge. In this paper, we introduce a novel framework for knowledge-based VQA titled “Prompting Large Language Models with Knowledge-Injection” (PLLMKI). We use vanilla VQA model to inspire the LLM and further enhance the LLM with knowledge injection. Unlike earlier approaches, we adopt the LLM for knowledge enhancement instead of relying on knowledge graphs. Furthermore, we leverage open LLMs, incurring no additional costs. In comparison to existing baselines, our approach exhibits the accuracy improvement of over 1.3 and 1.7 on two knowledge-based VQA datasets, namely OK-VQA and A-OKVQA, respectively.

Material extrusion additive manufacturing (AM) has gradually become a dominant technology for the fabrication of complex-designed thermoplastic polymers that require a higher level of control over the morphological and mechanical properties. The polymer internal crystal structure formed during the AM process can present significant impacts on the mechanical properties of the individual filaments, as well as the whole structure. Currently, limited details are known about the crystal structure evolution during the material extrusion AM processes of polymers. A novel in situ synchrotron X-ray diffraction (XRD) experimental configuration was developed enabling us to capture the material evolution data throughout the extrusion AM process. The in situ time-resolved data was analysed to reveal nucleation and crystallization sequences during the continuous deposition, with the aid of both complimentary numerical simulations and post-process (ex situ) characterisations. The thermal simulations supported the prediction of the filament temperature profile over time and location during the AM process, while ex situ characterisations validated the correlation between polymer crystallinity (resulting from printing parameters) and corresponding mechanical properties. The results obtained from varied process parameters suggest that the processing temperature has a dominant influence on the crystal microstructure evolution compared to the deposition velocity. A lower processing temperature just above the melting temperature permitted favourable crystallization conditions. The overall analysis demonstrated prospects for enhancing polymer AM, to engineering mechanically hierarchical structures through correlative investigations.

This study examined the effect of magneto-induced electric field (MIEF) treatment under varying pH conditions on the physicochemical, structural, and rheological properties of citrus pectin (CP). The results indicate that as the pH increased (to 4.0, 7.0, and 10.0), the molecular weight, degree of esterification, and galacturonic acid content of CP decreased. At pH values of 4.0 and 7.0, MIEF treatment caused fractures in the side chains of CP, while at pH 10.0, the main chain broke. MIEF treatment under these conditions enhanced the steady-state fluid behavior and dynamic viscoelastic properties of CP by promoting increased entanglements of neutral sugar side chains. Overall, the combined effect of pH and MIEF led to the breakdown of the main or side chains of CP, reducing its molecular weight and thereby influencing its rheological properties. In effect, this study provides a green modification technique for CP and its application to functional foods.

The use of properly designed nanostructured anode targets enables direct generation of high performance X-rays. In particular, this scheme has attracted widespread research interest for obtaining micro-focal spot X-ray radiation. In this study, we developed a thermal evaporating deposition technique and achieved controllable preparation of different single crystalline molybdenum (Mo) nanostructures by exploiting the lattice matching relationship between the substrate and the product crystal orientation. When used as an anode target for a reflective X-ray radiation device, these nanostructured Mo single crystals exhibit advantages over traditional targets at larger radiation angles. Further experimental and simulation results demonstrate the geometric characteristics of these Mo nanostructures and their correlation with the radiation intensity, focal spot size, and brightness of the generated X-rays. Among them, a unique frustum of square pyramid single crystalline Mo nanostructure stands out, due to its combination of three-dimensional body feature of nanostructures and radiation plane feature of planar targets. The research results not only provide a class of nanostructured anode target materials for reflective X-ray sources with different radiation angle requirements, but also make it possible to design and develop miniaturized devices for X-ray sources with selectable radiation performance and micro-focal spot.

Background: Immunotherapy favors patients with tumors; however, only 3–26.3% of patients with cervical cancer benefit from single-agent immune checkpoint inhibitors. Combined immunotherapy and chemotherapy has been explored against tumor; however, the combination remains controversial. This study aimed to investigate the tumor immune microenvironment (TIME) and the effects of platinum-based neoadjuvant chemotherapy (NACT) in cervical cancer to identify the clinical value of combining chemotherapy with immunotherapy. Methods: Multiplex immunohistochemistry (IHC) with 11 markers (cluster of differentiation [CD]3, CD8, CD4, CD11c, CD68, forkhead box P3 [Foxp3], programmed cell death 1 [PD-1], programmed cell death 1 ligand 1 [PD-L1], indoleamine 2,3-dioxygenase [IDO], cyclin-dependent kinase inhibitor 2A [p16], and cytokeratin [CK]) was performed to evaluate TIME from 108 matched pre- and post-NACT cervical cancer samples. The mechanism of antitumor immunity triggered by NACT was explored using RNA sequencing (RNA-seq) from four paired samples and subsequently verified in 41 samples using IHC. Results: The infiltration rate of the CD8+ T cells in treatment-naive cervical cancer was 0.73%, and those of Foxp3+ regulatory T cells (Tregs) and IDO+ cells were 0.87% and 17.15%, respectively. Moreover, immunoreactive T cells, dendritic cells, and macrophages were more in the stromal than the intratumor region. NACT increased dendritic, CD3+ T, CD8+ T, and CD4+ T cells and decreased Tregs. The aforementioned alterations occurred predominantly in the stromal region and were primarily in responders. Non-responders primarily showed decreased Tregs and no increase in CD8+ T or dendritic cell infiltration. Furthermore, dendritic cells interacted more closely with CD3+ T cells after NACT, an effect primarily observed in responders. RNA-seq data revealed activation of the antigen receptor-mediated signaling pathway and upregulation of major histocompatibility complex (MHC) I and MHC II after chemotherapy, validated using IHC. Conclusions: NACT can reduce Tregs, and when tumor cells are effectively killed, antigen presentation is enhanced, subsequently activating antitumor immunity finitely. Our study provides the molecular characteristics and theoretical basis for the simultaneous or sequential combination of platinum-based NACT and immunotherapy for cervical cancer.

Medical image segmentation plays a critical role in accurate diagnosis and treatment planning, enabling precise analysis across a wide range of clinical tasks. This review begins by offering a comprehensive overview of traditional segmentation techniques, including thresholding, edge-based methods, region-based approaches, clustering, and graph-based segmentation. While these methods are computationally efficient and interpretable, they often face significant challenges when applied to complex, noisy, or variable medical images. The central focus of this review is the transformative impact of deep learning on medical image segmentation. We delve into prominent deep learning architectures such as Convolutional Neural Networks (CNNs), Fully Convolutional Networks (FCNs), U-Net, Recurrent Neural Networks (RNNs), Adversarial Networks (GANs), and Autoencoders (AEs). Each architecture is analyzed in terms of its structural foundation and specific application to medical image segmentation, illustrating how these models have enhanced segmentation accuracy across various clinical contexts. Finally, the review examines the integration of deep learning with traditional segmentation methods, addressing the limitations of both approaches. These hybrid strategies offer improved segmentation performance, particularly in challenging scenarios involving weak edges, noise, or inconsistent intensities. By synthesizing recent advancements, this review provides a detailed resource for researchers and practitioners, offering valuable insights into the current landscape and future directions of medical image segmentation.

The complexity of urban road scenes at night and the inadequacy of visible light imaging in such conditions pose significant challenges. To address the issues of insufficient color information, texture detail, and low spatial resolution in infrared imagery, we propose an enhanced infrared detection model called LFIR-YOLO, which is built upon the YOLOv8 architecture. The primary goal is to improve the accuracy of infrared target detection in nighttime traffic scenarios while meeting practical deployment requirements. First, to address challenges such as limited contrast and occlusion noise in infrared images, the C2f module in the high-level backbone network is augmented with a Dilation-wise Residual (DWR) module, incorporating multi-scale infrared contextual information to enhance feature extraction capabilities. Secondly, at the neck of the network, a Content-guided Attention (CGA) mechanism is applied to fuse features and re-modulate both initial and advanced features, catering to the low signal-to-noise ratio and sparse detail features characteristic of infrared images. Third, a shared convolution strategy is employed in the detection head, replacing the decoupled head strategy and utilizing shared Detail Enhancement Convolution (DEConv) and Group Norm (GN) operations to achieve lightweight yet precise improvements. Finally, loss functions, PIoU v2 and Adaptive Threshold Focal Loss (ATFL), are integrated into the model to better decouple infrared targets from the background and to enhance convergence speed. The experimental results on the FLIR and multispectral datasets show that the proposed LFIR-YOLO model achieves an improvement in detection accuracy of 4.3% and 2.6%, respectively, compared to the YOLOv8 model. Furthermore, the model demonstrates a reduction in parameters and computational complexity by 15.5% and 34%, respectively, enhancing its suitability for real-time deployment on resource-constrained edge devices.

Abstract The emergence of large language models has led to the development of powerful tools such as ChatGPT that can produce text indistinguishable from human-generated work. With the increasing accessibility of such technology, students across the globe may utilize it to help with their school work—a possibility that has sparked ample discussion on the integrity of student evaluation processes in the age of artificial intelligence (AI). To date, it is unclear how such tools perform compared to students on university-level courses across various disciplines. Further, students’ perspectives regarding the use of such tools in school work, and educators’ perspectives on treating their use as plagiarism, remain unknown. Here, we compare the performance of the state-of-the-art tool, ChatGPT, against that of students on 32 university-level courses. We also assess the degree to which its use can be detected by two classifiers designed specifically for this purpose. Additionally, we conduct a global survey across five countries, as well as a more in-depth survey at the authors’ institution, to discern students’ and educators’ perceptions of ChatGPT’s use in school work. We find that ChatGPT’s performance is comparable, if not superior, to that of students in a multitude of courses. Moreover, current AI-text classifiers cannot reliably detect ChatGPT’s use in school work, due to both their propensity to classify human-written answers as AI-generated, as well as the relative ease with which AI-generated text can be edited to evade detection. Finally, there seems to be an emerging consensus among students to use the tool, and among educators to treat its use as plagiarism. Our findings offer insights that could guide policy discussions addressing the integration of artificial intelligence into educational frameworks.

Hydrogen is blended with natural gas to form hydrogenated natural gas (HCNG), which is a new, efficient and clean energy. CHEMKIN-PRO 19.0 software was combined with the GRI-Mech 3.0 mechanism to evaluate the capacity of H2 blending in reducing CO and CO2 emissions. Influences of H2 blending on combustion reactions of the CH4-air mixture were investigated. The results showed that the main reactants and products (CH4, CO, and CO2) decreased in gradient with increasing H2 blending ratio, accompanied by a shorter reaction duration and a faster reaction rate. After adding H2, important key radicals H, O, and OH increase significantly, so that the combustion reactions become more violent. Sensitivity analysis reveals that among relevant elementary reactions of CO and CO2, R38 (with its promotional effect) and R158 (with its inhibitory effect) show the greatest sensitivity. As the H2 concentration increases, the sensitivity of the two reactions (separately with promotional and inhibitory effects) decreases. Blending H2 in the natural gas can improve the combustion rate and reduce the generation of emissions CO and CO2, which is of important significance for realizing low-carbon goals and reducing air pollution.

Abstract Printing is a promising method for the large-scale, high-throughput, and low-cost fabrication of electronics. Specifically, the contact printing approach shows great potential for realizing high-performance electronics with aligned quasi-1D materials. Despite being known for more than a decade, reports on a precisely controlled system to carry out contact printing are rare and printed nanowires (NWs) suffer from issues such as location-to-location and batch-to-batch variations. To address this problem, we present here a novel design for a tailor-made contact printing system with highly accurate control of printing parameters (applied force: 0–6 N ± 0.3%, sliding velocity: 0–200 mm/s, sliding distance: 0–100 mm) to enable the uniform printing of nanowires (NWs) aligned along 93% of the large printed area (1 cm2). The system employs self-leveling platforms to achieve optimal alignment between substrates, whereas the fully automated process minimizes human-induced variation. The printing dynamics of the developed system are explored on both rigid and flexible substrates. The uniformity in printing is carefully examined by a series of scanning electron microscopy (SEM) images and by fabricating a 5 × 5 array of NW-based photodetectors. This work will pave the way for the future realization of highly uniform, large-area electronics based on printed NWs.

Abstract Pregnant women are generally more susceptible to viral infection. Although the impact of SARS-CoV-2 in pregnancy remains to be determined, evidence indicates that the risk factors for severe COVID-19 are similar in pregnancy to the general population. Here we systemically analyzed the clinical characteristics of pregnant and non-pregnant female COVID-19 patients who were hospitalized during the same period and found that pregnant patients developed marked lymphopenia and higher inflammation evident by higher C-reactive protein and IL-6. To elucidate the pathways that might contribute to immunopathology or protective immunity against COVID-19 during pregnancy, we applied single-cell mRNA sequencing to profile peripheral blood mononuclear cells from four pregnant and six non-pregnant female patients after recovery along with four pregnant and three non-pregnant healthy donors. We found normal clonal expansion of T cells in the pregnant patients, heightened activation and chemotaxis in NK, NKT, and MAIT cells, and differential interferon responses in the monocyte compartment. Our data present a unique feature in both innate and adaptive immune responses in pregnant patients recovered from COVID-19.

BackgroundHelicobacter pylori (HP), a gram-negative spiral-shaped microaerophilic bacterium, colonizes the stomach of approximately 50% of the world’s population, which is considered a risk factor for gastritis, peptic ulcers, gastric cancer, and other malignancies. HP is also considered carcinogenic since it involves the mutation and damage of multiple HP-related genes. Stomach adenocarcinoma (STAD) is a common stom5ach cancer with a poor prognosis and high risk of metastasis in the advanced stage. Therefore, an early diagnosis and targeted therapies are needed to ensure a better prognosis. In this study, a scoring system was constructed based on three HP infection–related candidate genes to enable a more accurate prediction of tumor progression and metastasis and response to immunotherapies.MethodsHP infection–induced mutation patterns of STAD samples from six cohorts were comprehensively assessed based on 73 HP-related genes, which were then correlated with the immune cell–infiltrating characteristics of the tumor microenvironment (TME). The risk signature was constructed to quantify the influence of HP infection on individual tumors. Subsequently, an accurate nomogram was generated to improve the clinical applicability of the risk signature. We conducted immunohistochemical experiments and used the Affiliated Hospital of Youjiang Medical University for Nationalities (AHYMUN) cohort data set with survival information to further verify the clinical value of this risk signature.ResultsTwo distinct HP-related mutation patterns with different immune cell–infiltrating characteristics (ICIC) and survival possibility were identified. We demonstrated that the evaluation of HP infection–induced mutation patterns of tumor could assist the prediction of stages, phenotypes, stromal activity, genetic diversity, and patient prognosis. A low risk score involved an increased mutation burden and activation of immune responses, with a higher 5-year survival rate and enhanced response to anti-PD-1/L1 immunotherapy, while a high risk score involved stromal activation and poorer survival. The efficiency of the risk signature was further evidenced by the nomogram.ConclusionsSTAD patients with a low risk score demonstrated significant therapeutic advantages and clinical benefits. HP infection–induced mutations play a nonnegligible role in STAD development. Quantifying the HP-related mutation patterns of individual tumors will contribute to phenotype classification, guide more effective targeted and personalized therapies, and enable more accurate predictions of metastasis and prognosis.

Objective Recent studies have reported a correlation between non-coding RNAs such as circular RNAs (circRNAs) and clinical value of various cancers. However, the diagnostic and prognostic role of circRNA in breast cancer remains controversial.Design Systematic review and meta-analysis.Methods Diagnostic efficacy was estimated by sensitivity, specificity and area under the curve (AUC). Pooled HRs with 95% CIs estimated overall survival (OS), and ORs with 95% CIs investigated clinical features.Results By searching PubMed, Embase, Web of Science, CNKI and Cochrane Library, we obtained a total of 29 studies with 4405 patients. A shorter survival time was associated with high expression levels of tumour-promoter circRNAs (OS: HR=2.43, 95% CI 2.20 to 2.92, p

In recent years, mobile phone location (MPL) data have been widely used to determine the spatial trajectories of users. Although this massive amount of MPL data can provide insight into human movement, definite conclusions cannot be drawn because of positioning bias: the locations of MPL data are usually not the phone users' actual locations. In recent years, the spatial accuracy of MPL data has been increasingly evaluated. Such efforts have led to many insights regarding the quality and applicability of MPL data. Despite these achievements, to the best of our knowledge, no studies have quantitatively assessed the spatial accuracy of MPL data by considering geographical influencing factors. In this study, we built a linear evaluation model based on geographical weighted regression (GWR) and a nonlinear evaluation model based on a random forest (RF) to quantify the relationship between geographical factors and the positioning bias of MPL data. Nanjing city in China is used as the test case. The results show that both the GWR model and RF model have good stability. However, the RF model's overall prediction performance is much better than that of the GWR model. The RF model can estimate the spatial accuracy of the MPL data within narrow margins of error. The importance ranking of geographical variables shows that the population density, elevation and building density are the three most important factors, while the normalised difference water index (NDWI) and distance to the nearest cell tower (DNCT) are the least important variables. The RF model constructed in this study can be used to evaluate the spatial accuracy of MPL data and simulate the spatial distribution of the positioning bias of the MPL data covering the study area.

Pregnant women are generally more susceptible to viral infection. Although the impact of SARS-CoV-2 on pregnant women remains to be determined, evidence indicates that risks of adverse clinical outcomes are similar in pregnancy to the general population. Here we analyzed clinical symptoms and outcomes of 20 pregnant and 299 reproductive-aged non-pregnant female COVID-19 patients who were hospitalized during the same period. Laboratory measurements were compared among mild cases and healthy pregnant women. Our study found that pregnant patients showed enhanced innate immune response evident by higher neutrophils and C-reactive protein. Cytokines, chemokines, and growth factors (CCGFs) profiles from 11 pregnant and 4 non-pregnant COVID-19 patients and 10 healthy pregnant female patients, and lymphocyte subsets analysis of 7 pregnant patients and 19 non-pregnant patients, indicate suppressed cytokine storm and potential enhanced CD8+ T cell and NK cell activity in pregnant patients with COVID-19, which may be essential in contributing to the unique anti-SARS-CoV-2 response in pregnancy.

Abstract This study investigates the potential interrelationship between gene expression and biosynthesis of vitamin C, E and folate in sweet corn sprouts. Germination of sweet corn kernels was conducted in light and dark environments to determine if this relationship was regulated by photo-illumination. Results indicated that light and dark environments affected the DHAR, TMT and GTPCH expression and that these genes were the predominant genes of vitamin C, E and folate biosynthesis pathways respectively during the germination. Levels of vitamin C and folate increased during the germination of sweet corn seeds while vitamin E had a declining manner. Sweet corn sprouts had higher vitamin C and E levels as well as relevant gene expression levels in light environment while illumination had little influence on the folate contents and the gene expression levels during the germination. These results indicate that there might be a collaborative relationship between vitamin C and folate regulation during sweet corn seed germination, while an inhibitive regulation might exist between vitamin C and E.

Electroactive biomaterials are fascinating for tissue engineering applications because of their ability to deliver electrical stimulation directly to cells, tissue, and organs. One particularly attractive conductive filler for electroactive biomaterials is silver nanoparticles (AgNPs) because of their high conductivity, antibacterial activity, and ability to promote bone healing. However, production of AgNPs involves a toxic reducing agent which would inhibit biological scaffold performance. This work explores facile and green synthesis of AgNPs using extract of Cilembu sweet potato and studies the effect of baking and precursor concentrations (1, 10 and 100 mM) on AgNPs’ properties. Transmission electron microscope (TEM) results revealed that the smallest particle size of AgNPs (9.95 ± 3.69 nm) with nodular morphology was obtained by utilization of baked extract and ten mM AgNO3. Polycaprolactone (PCL)/AgNPs scaffolds exhibited several enhancements compared to PCL scaffolds. Compressive strength was six times greater (3.88 ± 0.42 MPa), more hydrophilic (contact angle of 76.8 ± 1.7°), conductive (2.3 ± 0.5 × 10−3 S/cm) and exhibited anti-bacterial properties against Staphylococcus aureus ATCC3658 (99.5% reduction of surviving bacteria). Despite the promising results, further investigation on biological assessment is required to obtain comprehensive study of this scaffold. This green synthesis approach together with the use of 3D printing opens a new route to manufacture AgNPs-based electroactive with improved anti-bacterial properties without utilization of any toxic organic solvents.

Oxygen vacancies (OVs) have critical effects on the photoelectric characterizations and photocatalytic activity of nanoceria, but the contributions of surface OVs on the promoted photocatalytic properties are not clear yet. In this work, we synthesized ceria nanopolyhedron (P-CeO2), ceria nanocube (C-CeO2) and ceria nanorod (R-CeO2), respectively, and annealed them at 600 °C in air, 30%, 60% or pure H2. After annealing, the surface OVs concentration of ceria elevates with the rising of H2 concentration. Photocatalytic activity of annealed ceria is promoted with the increasing of surface OVs, the methylene blue photodegradation ratio with pure hydrogen annealed of P-CeO2, C-CeO2 or R-CeO2 is 93.82%, 85.15% and 90.09%, respectively. Band gap of annealed ceria expands first and then tends to narrow slightly with the rising of surface OVs, while the valence band (VB) and conductive band (CB) of annealed ceria changed slightly. Both of photoluminescence spectra and photocurrent results indicate that the separation efficiency of photoinduced electron-hole pairs is significantly enhanced with the increasing of the surface OVs concentration. The notable weakened recombination of photogenerated carrier is suggested to attribute a momentous contribution on the enhanced photocatalytic activity of ceria which contains surface OVs.

In this study, the phytochemical profiles, total and cellular antioxidant activities of five different Chinese chestnut (Castanea mollissima BL.) cultivars were analyzed. Phenolics, flavonoids as well as phytochemical compounds in five cultivars of chestnut kernels were determined. Results showed that the free forms played a dominant role in total phenolics, flavonoids and antioxidant activities of all five cultivars of chestnut kernels. The cultivar ‘Fyou’ showed the highest total and free phenolic contents, ‘Heguoyihao’ showed the highest total and free flavonoids contents, and ‘Chushuhong’ showed the highest total and cellular antioxidant activities. Eight phenolic compounds were detected, and chlorogenic acid, gallic acid, and quercetin were shown as three predominant components in all five cultivars. These results provide valuable information which may be a guidance for selection of good chestnut variety to be used as functional food.

The intention of this study was to investigate the effect of steaming processing on phenolic profiles and antioxidant activities in chestnuts. Steaming processing at different times and temperatures depicted diverse impacts on free and bound fractions. Though, bound phenolics were stable but long time steaming at higher temperatures tended to improve the levels of phenolics, flavonoids as well as antioxidant activities in chestnut kernels, by up to 60.11% of the original value. Seven phenolic compounds including ferulic acid, chlorogenic acid, gallic acid, vanillic acid, syringate, p-coumaric acid and quercetin were found to change during thermal processes. Significant relationships (p < 0.05) were identified between total phenolics and total antioxidant activities. However, the consistency of chlorogenic acid (p < 0.01) only with cellular antioxidant activity indicated poor bio-accessibility of the phytochemicals in chestnuts. However, this situation could be partly improved by steaming. Steaming could improve the cellular accessibility of free phytochemicals, particularly, increasing the bio-accessibility by 41.96%. This study provided valuable information on dynamic changes of phenolic profiles and antioxidant activity of chestnuts under a steaming process, which could offer possible guidance for the chestnut processing industry in the future.