Lightweight Vehicle Detection Based on Improved YOLOv5s.

A vehicle detection algorithm is an indispensable component of intelligent traffic management and control systems, influencing the efficiency and functionality of the system. In this paper, we propose a lightweight improvement method for the YOLOv5 algorithm based on integrated perceptual attention, with few parameters and high detection accuracy. First, we propose a lightweight module IPA with a Transformer encoder based on integrated perceptual attention, which leads to a reduction in the number of parameters while capturing global dependencies for richer contextual information. Second, we propose a lightweight and efficient multiscale spatial channel reconstruction (MSCCR) module that does not increase parameter and computational complexity and facilitates representative feature learning. Finally, we incorporate the IPA module and the MSCCR module into the YOLOv5s backbone network to reduce model parameters and improve accuracy. The test results show that, compared with the original model, the model parameters decrease by about 9%, the average accuracy (mAP@50) increases by 3.1%, and the FLOPS does not increase.

KRP-DS: A Knowledge Graph-Based Dialogue System with Inference-Aided Prediction.

With the popularity of ChatGPT, there has been increasing attention towards dialogue systems. Researchers are dedicated to designing a knowledgeable model that can engage in conversations like humans. Traditional seq2seq dialogue models often suffer from limited performance and the issue of generating safe responses. In recent years, large-scale pretrained language models have demonstrated their powerful capabilities across various domains. Many studies have leveraged these pretrained models for dialogue tasks to address concerns such as safe response generation. Pretrained models can enhance responses by carrying certain knowledge information after being pre-trained on large-scale data. However, when specific knowledge is required in a particular domain, the model may still generate bland or inappropriate responses, and the interpretability of such models is poor. Therefore, in this paper, we propose the KRP-DS model. We design a knowledge module that incorporates a knowledge graph as external knowledge in the dialogue system. The module utilizes contextual information for path reasoning and guides knowledge prediction. Finally, the predicted knowledge is used to enhance response generation. Experimental results show that our proposed model can effectively improve the quality and diversity of responses while having better interpretability, and outperforms baseline models in both automatic and human evaluations.

EFC-YOLO: An Efficient Surface-Defect-Detection Algorithm for Steel Strips.

The pursuit of higher recognition accuracy and speed with smaller model sizes has been a major research topic in the detection of surface defects in steel. In this paper, we propose an improved high-speed and high-precision Efficient Fusion Coordination network (EFC-YOLO) without increasing the model's size. Since modifications to enhance feature extraction in shallow networks tend to affect the speed of model inference, in order to simultaneously ensure the accuracy and speed of detection, we add the improved Fusion-Faster module to the backbone network of YOLOv7. Partial Convolution (PConv) serves as the basic operator of the module, which strengthens the feature-extraction ability of shallow networks while maintaining speed. Additionally, we incorporate the Shortcut Coordinate Attention (SCA) mechanism to better capture the location information dependency, considering both lightweight design and accuracy. The de-weighted Bi-directional Feature Pyramid Network (BiFPN) structure used in the neck part of the network improves the original Path Aggregation Network (PANet)-like structure by adding step branches and reducing computations, achieving better feature fusion. In the experiments conducted on the NEU-DET dataset, the final model achieved an 85.9% mAP and decreased the GFLOPs by 60%, effectively balancing the model's size with the accuracy and speed of detection.

Supplementation of citrus pectin with whole-cell pectinase PG5 on Pichia pastoris promotes recovery of colitis and enhances intestinal barrier function in DSS-treated mice.

A whole-cell biocatalyst was developed by genetically engineering pectinase PG5 onto the cell surface of Pichia pastoris using Gcw12 as the anchoring protein. Whole-cell PG5 eliminated the need for enzyme extraction and purification, while also exhibiting enhanced thermal stability, pH stability, and resistance to proteases in vitro compared to free PG5. Magnetic resonance mass spectrometry analysis revealed that whole-cell PG5 efficiently degraded citrus pectin, resulting in the production of a mixture of pectin oligosaccharides. The primary components of the mixture were trigalacturonic acid, followed by digalacturonic acid and tetragalacturonic acid. Supplementation of citrus pectin with whole-cell PG5 resulted in a more pronounced protective effect compared to free PG5 in alleviating colitis symptoms and promoting the integrity of the colonic epithelial barrier in a mouse model of dextran sulfate sodium-induced colitis. Hence, this study demonstrates the potential of utilizing whole-cell pectinase as an effective biocatalyst to promote intestinal homeostasis in vivo.

Stillbirths attributable to open fires and their geographic disparities in non-Western countries.

Due to global warming, an increased number of open fires is becoming a major contributor to PM2.5 pollution and thus a threat to public health. However, the burden of stillbirths attributable to fire-sourced PM2.5 is unknown. In low- and middle-income countries (LMICs), there is a co-occurrence of high baseline stillbirth rates and frequent firestorms, which may lead to a geographic disparity. Across 54 LMICs, we conducted a self-matched case-control study, making stillbirths comparable to the corresponding livebirths in terms of time-invariant characteristics (e.g., genetics) and duration of gestational exposure. We established a joint-exposure-response function (JERF) by simultaneously associating stillbirth with fire- and non-fire-sourced PM2.5 concentrations, which were estimated by fusing multi-source data, such as chemical transport model simulations and satellite observations. During 2000-2014, 35,590 pregnancies were selected from multiple Demographic and Health Surveys. In each mother, a case of stillbirth was compared to her livebirth(s) based on gestational exposure to fire-sourced PM2.5. We further applied the JERF to assess stillbirths attributable to fire-sourced PM2.5 in 136 non-Western countries. The disparity was evaluated using the Gini index. The risk of stillbirth increased by 17.4% (95% confidence interval [CI]: 1.6-35.7%) per 10 µg/m3 increase in fire-sourced PM2.5. In 2014, referring to a minimum-risk exposure level of 10 µg/m3, total and fire-sourced PM2.5 contributed to 922,860 (95% CI: 578,451-1,183,720) and 49,951 (95% CI: 3,634-92,629) stillbirths, of which 10% were clustered within the 6.4% and 0.6% highest-exposure pregnancies, respectively. The Gini index of stillbirths attributable to fire-sourced PM2.5 was 0.65, much higher than for total PM2.5 (0.28). Protecting pregnant women against PM2.5 exposure during wildfires is critical to avoid stillbirths, as the burden of fire-associated stillbirths leads to a geographic disparity in maternal health.

Nickel-catalysed cross-electrophile coupling of aryl bromides and primary alkyl bromides.

The structure of primary alkylated arenes plays an important role in the molecular action of drugs and natural products. The nickel/spiro-bidentate-pyox catalysed cross-electrophile coupling of aryl bromides and primary alkyl bromides was developed for the formation of the Csp2-Csp3 bond, which provided an efficient method for the synthesis of primary alkylated arenes. The reactions could tolerate functional groups such as ester, aldehyde, ketone, ether, benzyl, and imide.

Estimation of stillbirths attributable to ambient fine particles in 137 countries.

Gestational exposure to ambient fine particles (PM2.5) increases the risk of stillbirth, but the related disease burden is unknown, particularly in low- and middle-income countries (LMICs). We combine state-of-the-art estimates on stillbirths, and multiple exposure-response functions obtained from previous meta-analyses or derived by a self-matched case-control study in 54 LMICs. 13,870 stillbirths and 32,449 livebirths are extracted from 113 geocoded surveys from the Demographic and Health Surveys. Each stillbirth is compared to livebirth(s) of the same mother using a conditional logit regression. We find that 10-µg/m3 increase of PM2.5 is associated with an 11.0% (95% confidence interval [CI] 6.4, 15.7) increase in the risk of stillbirth, and the association is significantly enhanced by maternal age. Based on age-specific nonlinear PM2.5-stillbirth curves, we evaluate the PM2.5-related stillbirths in 137 countries. In 2015, of 2.09 (95% CI: 1.98, 2.20) million stillbirths, 0.83 (0.54, 1.08) million or 39.7% (26.1, 50.8) are attributable to PM2.5 exposure exceeding the reference level of 10 µg/m3. In LMICs, preventing pregnant women from being exposed to PM2.5 can improve maternal health.