Search

The ability to predict city-wide parking availability is crucial for the successful development of Parking Guidance and Information (PGI) systems. Indeed, the effective prediction of city-wide parking availability can improve parking efficiency, help urban planning and ultimately alleviate city congestion. However, it is a non-trivial task because of three major challenges:1) the non-Euclidean spatial autocorrelation among parking lots, 2) the dynamic temporal autocorrelation inside of and between parking lots, and 3) the scarcity of real-time parking availability information obtained from real-time sensors. To this end, we propose Semi-supervised Hierarchical Recurrent Graph Neural Network-X(SHARE-X) for predicting city-wide parking availability. Specifically, we first propose a hierarchical graph convolution structure to model non-Euclidean spatial autocorrelation among parking lots. Along this line, a contextual graph convolution block and a multi-resolution soft clustering graph convolution block are respectively proposed to capture local and global spatial dependencies between parking lots. Moreover, we devise a hierarchical attentive recurrent network module to incorporate both short and long-term dynamic temporal dependencies of parking lots. Additionally, a parking availability approximation module is introduced to estimate missing real-time parking availabilities from both spatial and temporal domains. Finally,experiments on two real-world datasets demonstrate the prediction performance of SHARE-X outperforms eight state-of-the-art baselines.

Modeling the global dynamics of emerging infectious diseases (EIDs) like COVID-19 can provide important guidance in the preparation and mitigation of pandemic threats. While age-structured transmission models are widely used to simulate the evolution of EIDs, most of these studies focus on the analysis of specific countries and fail to characterize the spatial spread of EIDs across the world. Here, we developed a global pandemic simulator that integrates age-structured disease transmission models across 3,157 cities and explored its usage under several scenarios. We found that without mitigations, EIDs like COVID-19 are highly likely to cause profound global impacts. For pandemics seeded in most cities, the impacts are equally severe by the end of the first year. The result highlights the urgent need for strengthening global infectious disease monitoring capacity to provide early warnings of future outbreaks. Additionally, we found that the global mitigation efforts could be easily hampered if developed countries or countries near the seed origin take no control. The result indicates that successful pandemic mitigations require collective efforts across countries. The role of developed countries is vitally important as their passive responses may significantly impact other countries.

User identification across multiple datasets has a wide range of applications and there has been an increasing set of research works on this topic during recent years. However, most of existing works focus on user identification with a single input data type, e.g., (I) identifying a user across multiple social networks with online data and (II) detecting a single user from heterogeneous trajectory datasets with offline data. Different from previous works, in this paper, we propose a framework on user identification between online and offline datasets. We build connections between these two types of data by a mapping from IP addresses to physical locations. To solve this problem, we propose a novel framework consists of three steps. First, we use a clustering method based on locations of IP addresses to map IP addresses into specific physical location distributions. Second, we propose a novel pairwise index to reduce space cost and running time for computing the co-occurrence. Lastly, we apply a learning-to-rank method to merge the effect of multiple features we get in the first two steps. Based on our framework, we design experiments to demonstrate the efficiency (in time and space) of our framework, together with the precision and recall of our approach compared to other methods.

The transmission of certain bacteria from the mother's gut to the infant's gut via breast milk (BM) is critical for offspring's immune system development. Dysbiosis of the BM microbiota can be caused by a variety of reasons, which can be influenced by probiotics delivered via the enteromammary route. The goal of this study was to investigate the bacterium that can be transmitted from mother to infant's intestine during various lactation periods in 19 mother-child dyads. Bacterial transmission is most common during the colostrum phase when bacteria with certain amplicon sequence variants (ASVs) enter the newborn intestine and inhabit it permanently. We have established that anaerobic gut-associated bacteria, such as Faecalibacterium, Blautia and Lachnoclostridium, transfer from the mother to the infant's gut with lactation dependence using the idea of weighted transfer ratios. Streptococcus salivarius, Bifidobacterium longum, and Lactobacillus gasseri were transferred from the maternal gut to the BM, as well as from the BM to the newborn gut, dependent on different ASVs. These findings suggest that isolating key microorganisms from breast milk could be utilized to modify the microbiota of BM or newborns by giving the mother a probiotic or adding it to artificial milk to promote neonatal health.

The consumption of mixed nuts is a healthy dietary strategy to reduce the risk of cardiovascular disease and has a prebiotic effect on the gut microbiota. However, there is a lack of basic research based on mixed nut formulation. This study established a new method for optimizing mixed nut formulations using the Nutrient Rich Food (NRF) index model. Nutrient indices were adjusted by combining 10 and 8 encouraging nutrients and 3 limiting nutrients of nuts and dried fruits, respectively. The optimized mixed nut formulation had the highest total NRF and the lowest energy, which was achieved by applying linear programming. The effect of an optimized mixed nut formulation on insulin resistance and gut microbiota was investigated in an animal model of metabolic disorders caused by a high-fat diet. Male C57BL/6J mice (

Gut microbiota transmission from mother to offspring is critical to infant gut microbiota and immune development. Mother's intestine and breast milk are rich in secretory immunoglobulin A (sIgA), which can coat a specific bacterial spectrum and may be related to bacterial transmission and colonization. Here we analyzed the microbiota and sIgA-coated bacteria of maternal fecal samples and breast milk and infant fecal samples from 19 dyads by 16S rRNA amplicon sequencing. For the sIgA-coated microbiota, both the phylogenetic diversity and the Shannon index of maternal fecal samples show a lower trend than those of infant fecal samples (

Different nanostructured CoP microcubes derived from metal formate frameworks with enhanced oxygen evolution reaction performance.

China, with its growing population and economic development, faces increasing risks to health from climate change, but also opportunities to address these risks and protect health for generations to come. Without a timely and adequate response, climate change will impact lives and livelihoods at an accelerated rate. In 2020, the Lancet Countdown Regional Centre in Asia, led by Tsinghua University, built on the work of the global Lancet Countdown and began its assessment of the health profile of climate change in China with the aim of triggering rapid and health-responsive actions. This 2021 report is the first annual update, presenting 25 indicators within five domains: climate change impacts, exposures, and vulnerability; adaptation, planning, and resilience for health; mitigation actions and health co-benefits; economics and finance; and public and political engagement. The report represents the contributions of 88 experts from 25 leading institutions in, and outside of, China. From 2020 to 2021, five new indicators have been added and methods have been improved for many indicators. Where possible, the indicator results are presented at national and provincial levels to facilitate local understanding and policy making. In a year marked by COVID-19, this report also endeavours to reflect on China's pathway for a green recovery, ensuring it aligns with the carbon neutrality goal, for the health of the current and future generations.

Left unmitigated, climate change poses a catastrophic risk to human health, demanding an urgent and concerted response from every country. The 2015 Lancet Commission on Health and Climate Change and The Lancet Countdown: Tracking Progress on Health and Climate Change have been initiated to map out the impacts of climate change and the necessary policy responses. To meet these challenges, Tsinghua University, partnering with the University College London and 17 Chinese and international institutions, has prepared the Chinese Lancet Countdown report, which has a national focus and builds on the work of the global Lancet Countdown: Tracking Progress on Health and Climate Change. Drawing on international methodologies and frameworks, this report aims to deepen the understanding of the links between public health and climate change at the national level and track them with 23 indicators. This work is part of the Lancet's Countdown broader efforts to develop regional expertise on this topic, and coincides with the launch of the Lancet Countdown Regional Centre in Asia, based at Tsinghua University. The data and results of this report are presented at the provincial level, where possible, to facilitate targeted response strategies for local decision-makers. Based on the data and findings of the 2020 Chinese Lancet Countdown report, five recommendations are proposed to key stakeholders in health and climate change in China: (1) Enhance inter-departmental cooperation. Climate change is a challenge that demands an integrated response from all sectors, urgently requiring substantial inter-departmental cooperation among health, environment, energy, economic, financial, and education authorities. (2) Strengthen health emergency preparedness. Knowledge and findings on current and future climate-related health threats still lack the required attention and should be fully integrated into the emergency preparedness and response system. (3) Support research and raise awareness. Additional financial support should be allocated to health and climate change research in China to enhance health system adaptation, mitigation measures, and their health benefits. At the same time, media and academia should be fully motivated to raise the public and politicians' awareness of this topic. (4) Increase climate change mitigation. Speeding up the phasing out of coal is necessary to be consistent with China's pledge to be carbon neutral by 2060 and to continue to reduce air pollution. Fossil fuel subsidies must also be phased out. (5) Ensure the recovery from COVID-19 to protect health now and in the future. China's efforts to recover from COVID-19 will shape public health for years to come. Climate change should be a priority in these interventions.

Recent years have witnessed a large amount of decentralized data in various (edge) devices of end-users, while the decentralized data aggregation remains complicated for machine learning jobs because of regulations and laws. As a practical approach to handling decentralized data, Federated Learning (FL) enables collaborative global machine learning model training without sharing sensitive raw data. The servers schedule devices to jobs within the training process of FL. In contrast, device scheduling with multiple jobs in FL remains a critical and open problem. In this paper, we propose a novel multi-job FL framework, which enables the training process of multiple jobs in parallel. The multi-job FL framework is composed of a system model and a scheduling method. The system model enables a parallel training process of multiple jobs, with a cost model based on the data fairness and the training time of diverse devices during the parallel training process. We propose a novel intelligent scheduling approach based on multiple scheduling methods, including an original reinforcement learning-based scheduling method and an original Bayesian optimization-based scheduling method, which corresponds to a small cost while scheduling devices to multiple jobs. We conduct extensive experimentation with diverse jobs and datasets. The experimental results reveal that our proposed approaches significantly outperform baseline approaches in terms of training time (up to 12.73 times faster) and accuracy (up to 46.4% higher)., To appear in TPDS; 22 pages, 17 figures, 8 tables. arXiv admin note: substantial text overlap with arXiv:2112.05928

Diabetic nephropathy is a serious complication of diabetes. Hyperoside has been widely reported to ameliorate diabetes-associated disease. The current study is designed to explore the mechanism of hyperoside in diabetic nephropathy. In the present study, high glucose was used to treat podocytes. Diabetic nephropathy mice models were established by high-fat feeding followed by multiple low dose injections of streptozocin. Western blot analysis was conducted for detection of extracellular matrix accumulation, inflammatory response and cell apoptosis. We found out that hyperoside improved high glucose-induced cell injury. Additionally, hyperoside prevented mice with diabetic nephropathy from diabetic symptoms and renal dysfunction. Mechanistically, hyperoside inhibited the mRNA and protein expression of APC. MiR-499-5p was found to be an upstream negative mediator of APC, and hyperoside induced the upregulation of miR-499-5p. MiR-499-5p bound with the 3' untranslated region of APC to inhibit its expression. Finally, rescue assays revealed that the suppressive effects of miR-499-5p overexpression on renal dysfunction were rescued by upregulation of APC in mice with diabetic nephropathy. In conclusion, these findings indicated that hyperoside ameliorates diabetic nephropathy via targeting the miR-499-5p/APC axis, suggesting that hyperoside may offer a potential tactic for diabetic nephropathy treatment.

Changes in secretory immunoglobulin A (SIgA) coated bacteria from early to late pregnancy were associated with the development of gestational diabetes mellitus (GDM). SIgA coated beneficial gut bacteria, which are depleted in GDM, are potential probiotics for the prevention of GDM. We investigated blood biochemistry, chronic inflammation, mucosal barrier biomarkers and faecal SIgA coated microbiota in healthy early pregnancy (T1H, n = 50), late pregnancy (T3H, n = 30) and women with GDM (T3D, n = 27). The “leaky gut” markers, zonulin and lipopolysaccharide (LPS), significantly increased in T3D compared to the T3H group. The Shannon index of SIgA coated microbiota was elevated in late pregnancy compared to early pregnancy and was the highest in the T3D group (p < 0.001). The T3D group was enriched in SIgA coated Escherichia and Streptococcus and depleted in Lactobacillus and Bifidobacterium. Blood glucose (BG) positively correlated with zonulin (p < 0.001) and LPS (p < 0.05). Lactobacillus reuteri negatively correlated with BG (p < 0.05), zonulin (p < 0.05) and LPS (p < 0.01). Lactobacillus reuteri QS01 isolated from the feces of T1H significantly reduced LPS released by the gut microbiota of GDM individuals in vitro. In conclusion, GDM may be related to intestinal mucosal damage and inflammation-induced dysbiosis of SIgA coated microbiota. SIgA coated L. reuteri can reduce the level of LPS of GDM in vitro.

Line structured light sensor has been applied in various three dimensional (3D) measurement scenes with the advantages of non-contact, low cost and high speed. Its accuracy, which is directly determined by the calibration method, needs to be further improved to fulfill the measuring tasks of precision parts. Here, we proposed a numerical method that can eliminate the errors of the model based methods through two strategies. One is the establishment of the numerical mapping relationship between stripe pixel coordinates and their world coordinates through piecewise cubic interpolation. Corner points of a checkerboard target are used to obtain sufficient interpolating nodes. This target can be manually aligned with the laser plane and the alignment error would be eliminated via the point projection. The other is the data-driven laser plane optimization. The data set of reference interval distance is computed based on the invariance of cross ratio. The optimization model is to minimize the root mean squared error of measured interval distances by adjusting the laser plane coefficients. After the optimization, a higher numerical mapping relationship can be achieved. It bypasses the camera and the distortion models and reaches a calibration error of only 0.005mm. The comparison studies and the measurement of the steps further validate the proposed method.

Considering the rigid-flexible coupling system of 5 MW wind turbine, the static pressure, fluid temperature, Mach number, vorticity, shear stress and velocity distribution of tower flange are obtained under the combined action of wind shear and fluctuating wind. The dynamic response of rigid-flexible coupling structure is studied by the combined action of the whole vibration and local vibration of 5 MW wind turbine. Through the analysis of dynamic parameter change curve, the improvement measures of tower structure are put forward, and the dynamic response of the structure before and after improvement is compared. The superiority of the improved structure is expounded, so as to provide reference basis for engineering application.

Human mobility annotation aims to assign mobility records the corresponding visiting Point-of-Interests (POIs). It is one of the most fundamental problems for understanding human mobile behaviors. In literature, many efforts have been devoted to annotating mobility records in a pointwise or trajectory-wise manner. However, the user preference factor is not fully explored and, worse still, the mobile peer influence factor has never been integrated. To this end, in this article, we propose a novel framework, named JEPPI, to jointly exploit user preference and mobile peer influence to tackle the problem. In our JEPPI, we first unify the two distinct factors in a behavior-driven user-POI graph. This graph enables us to model user preference with user-POI visiting relationships, and model two types of mobile peer influence with co-location and co-visiting peer relationships, respectively. Moreover, we devise an equivalence-emphasizing metric to reduce redundancy in the second-order co-visiting peer influence. In addition, a mutual augmentation learning approach is proposed to preserve the latent structures of various factors exploited. Notably, our learning approach preserves all factors in a shared representation space such that user preference is learned with mobile peer influence being considered at the same time, and vice versa. In this way, the different factors are mutually augmented and semantically integrated to enhance human mobility annotation. Finally, using two large-scale real-world datasets, we conduct extensive experiments to demonstrate the superiority of our approach compared with the state-of-the-art annotation methods.

The ability to predict city-wide parking availability is crucial for the successful development of Parking Guidance and Information (PGI) systems. Indeed, the effective prediction of city-wide parking availability can improve parking efficiency, help urban planning, and ultimately alleviate city congestion. However, it is a non-trivial task for predicting citywide parking availability because of three major challenges: 1) the non-Euclidean spatial autocorrelation among parking lots, 2) the dynamic temporal autocorrelation inside of and between parking lots, and 3) the scarcity of information about real-time parking availability obtained from real-time sensors (e.g., camera, ultrasonic sensor, and GPS). To this end, we propose Semi-supervised Hierarchical Recurrent Graph Neural Network (SHARE) for predicting city-wide parking availability. Specifically, we first propose a hierarchical graph convolution structure to model non-Euclidean spatial autocorrelation among parking lots. Along this line, a contextual graph convolution block and a soft clustering graph convolution block are respectively proposed to capture local and global spatial dependencies between parking lots. Additionally, we adopt a recurrent neural network to incorporate dynamic temporal dependencies of parking lots. Moreover, we propose a parking availability approximation module to estimate missing real-time parking availabilities from both spatial and temporal domain. Finally, experiments on two real-world datasets demonstrate the prediction performance of SHARE outperforms seven state-of-the-art baselines., Comment: 8 pages, 9 figures, AAAI-2020

Automatic diagnosis systems have attracted increasing attention in recent years. The reinforcement learning (RL) is an attractive technique for building an automatic diagnosis system due to its advantages for handling sequential decision making problem. However, the RL method still cannot achieve good enough prediction accuracy. In this paper, we propose a Generative Adversarial regularized Mutual information Policy gradient framework (GAMP) for automatic diagnosis which aims to make a diagnosis rapidly and accurately. We first propose a new policy gradient framework based on the Generative Adversarial Network (GAN) to optimize the RL model for automatic diagnosis. In our framework, we take the generator of GAN as a policy network, and also use the discriminator of GAN as a part of the reward function. This generative adversarial regularized policy gradient framework can try to avoid generating randomized trials of symptom inquires deviated from the common diagnosis paradigm. In addition, we add mutual information to enhance the reward function to encourage the model to select the most discriminative symptoms to make a diagnosis. Experiment evaluations on two public datasets show that our method beats the state-of-art methods, not only can achieve higher diagnosis accuracy, but also can use a smaller number of inquires to make diagnosis decision.

Purpose. To probe the property of serum uric acid in evaluating the activity of patients with thyroid-associated ophthalmopathy. Methods. A total of 443 patients with TAO admitted to our hospital from March 2016 to February 2021 were selected for the observation group. Simultaneously, 443 healthy subjects were selected for the control group. The observation group was divided into the active group ( n = 254 ) and the inactive group ( n = 189 ) according to the clinical activity score (CAS). Besides, the patients were divided into mild group ( n = 201 ), moderate severe group ( n = 133 ) and extremely severe group ( n = 109 ) based on the severity of TAO. Serum uric acid, free triiodothyronine (FT3), free thyroid hormone (FT4), thyrotropin stimulating hormone (TSH) and glycosylated hemoglobin (HbA1c) levels were detected and analyzed in each group. Results. Serum UA, FT3, FT4, TSH and HbA1c in the active group were significantly enhanced than those in the other two groups ( P < 0.05 ), and there was no significant difference between the inactive group and the control group ( P > 0.05 ). In different disease severity groups, the serum UA level of patients in the active group was significantly promoted than that in the inactive group and control group ( P < 0.05 ) and was decreased successively in extremely severe group, moderate severe group and mild group, with statistical significance ( P < 0.05 ). Pearson’s analysis showed that UA was positively correlated with FT3, FT4, and HbA1c ( r = 0.652 , P = 0.031 ; r = 0.571 , P = 0.042 ; r = 0.737 , P = 0.024 ), while was reversely correlated with TSH level ( r = − 0.137 , P = 0.262 ). There was no correlation between UA and FT3, FT4, and HbA1c levels in the inactive group. UA detection showed the average sensitivity and specificity of TAO activity were 94.3% and 85.2%, respectively. There was no significant correlation between the severity of disease and serum UA in inactive patients ( P = 0.135 ). There was a positive correlation between the severity of disease and serum UA in active patients ( P = 0.005 ). Conclusion. UA may be used as a laboratory indicator for quantitative clinical diagnosis of thyroid-associated ophthalmopathy (TAO) and as a parameter for the presence of TAO activity.

While online advertising is one of the major sources of income for search engines, pumping up the incomes from business advertisements while ensuring the user experience becomes a challenging but emerging area. Designing high-quality advertisements with persuasive content has been proved as a way to increase revenues through improving the Click-Through Rate (CTR). However, it is difficult to scale up the design of high-quality ads, due to the lack of automation in creativity. In this paper, we present Commonsense-Enriched Advertisement on Search Engine (CHASE) --- a system for the automatic generation of persuasive ads. CHASE adopts a specially designed language model that fuses the keywords, commonsense-related texts, and marketing contents to generate persuasive advertisements. Specifically, the language model has been pre-trained using massive contents of explicit knowledge and fine-tuned with well-constructed quasi-parallel corpora with effective control of the proportion of commonsense in the generated ads and fitness to the ads' keywords. The effectiveness of the proposed method CHASE has been verified by real-world web traffics for search and manual evaluation. In A/B tests, the advertisements generated by CHASE would bring 11.13% CTR improvement. The proposed model has been deployed to cover three advertisement domains (which are kid education, psychological counseling, and beauty e-commerce) at Baidu, the world's largest Chinese search engine, with adding revenue of about 1 million RMB (Chinese Yuan) per day.