Search

Geospatial artificial intelligence (GeoAI) has emerged as a subfield of GIScience that uses artificial intelligence approaches and machine learning techniques for geographic knowledge discovery. The non-regularity of data structures has recently led to different variants of graph neural networks in the field of computer science, with graph convolutional neural networks being one of the most prominent that operate on non-euclidean structured data where the numbers of nodes connections vary and the nodes are unordered. These networks use graph convolution – commonly known as filters or kernels – in place of general matrix multiplication in at least one of their layers. This paper suggests spatial regression graph convolutional neural networks (SRGCNNs) as a deep learning paradigm that is capable of handling a wide range of geographical tasks where multivariate spatial data needs modeling and prediction. The feasibility of SRGCNNs lies in the feature propagation mechanisms, the spatial locality nature, and a semi-supervised training strategy. In the experiments, this paper demonstrates the operation of SRGCNNs with social media check-in data in Beijing and house price data in San Diego. The results indicate that a well-trained SRGCNN model is capable of learning from samples and performing reasonable predictions for unobserved locations. The paper also presents the effectiveness of incorporating the idea of geographically weighted regression for handling heterogeneity between locations in the model approach. Compared to conventional spatial regression approaches, SRGCNN-based models tend to generate much more accurate and stable results, especially when the sampling ratio is low. This study offers to bridge the methodological gap between graph deep learning and spatial regression analytics. The proposed idea serves as an example to illustrate how spatial analytics can be combined with state-of-the-art deep learning models, and to enlighten future research at the front of GeoAI. [ABSTRACT FROM AUTHOR]

Policy decision of curling refers to providing strategy suggestions for curling competition with the help of computers. Existing curling agents have achieved good results in the digital scenarios, but cannot make correct decisions when applied to actual competition and training scenes. In this paper, a strategies decision agent in the real scene has been proposed. The competition situation was acquired by a Situation-Aware Network and mapped by a Digital Extraction module. We designed Curling MCTS to explore the best strategy in continuous space. The effectiveness of our framework has been verified by experiments and evaluated by China's wheelchair curling team at China Disabled Sports Management Center. With the help of our system, China's wheelchair curling team trained effectively and won the championship in the XIII Paralympic Winter Games (2022, Beijing). In addition, a new curling target detection dataset was provided. [ABSTRACT FROM AUTHOR]

Long-term exposure to air environments full of suspended particles, especially PM2.5, would seriously damage people's health and life (i.e., respiratory diseases and lung cancers). Therefore, accurate PM2.5 prediction is important for the government authorities to take preventive measures. In this paper, the advantages of convolutional neural networks (CNN) and long short-term memory networks (LSTM) models are combined. Then a hybrid CNN-LSTM model is proposed to predict the daily PM2.5 concentration in Beijing based on spatiotemporal correlation. Specifically, a Pearson's correlation coefficient is adopted to measure the relationship between PM2.5 in Beijing and air pollutants in its surrounding cities. In the hybrid CNN-LSTM model, the CNN model is used to learn spatial features, while the LSTM model is used to extract the temporal information. In order to evaluate the proposed model, three evaluation indexes are introduced, including root mean square error, mean absolute percent error, and R-squared. As a result, the hybrid CNN-LSTM model achieves the best performance compared with the Multilayer perceptron model (MLP) and LSTM. Moreover, the prediction accuracy of the proposed model considering spatiotemporal correlation outperforms the same model without spatiotemporal correlation. Therefore, the hybrid CNN-LSTM model can be adopted for PM2.5 concentration prediction. [ABSTRACT FROM AUTHOR]

Finding students at high risk of poor academic performance as early as possible plays an important role in improving education quality. To do so, most existing studies have used the traditional machine learning algorithms to predict students' achievement based on their behavior data, from which behavior features are extracted manually thanks to expert experience and knowledge. However, owing to an increase in the varieties and overall volume of behavioral data, it has become more and more challenging to identify high-quality handcrafted features. In this paper, we propose an end-to-end deep learning model that automatically extracts features from students' multi-source heterogeneous behavior data to predict academic performance. The key innovation of this model is that it uses long short-term memory networks to capture inherent time-series features for each type of behavior, and it takes two-dimensional convolutional networks to extract correlation features among different behaviors. We conducted experiments with four types of daily behavior data from students of the university in Beijing. The experimental results demonstrate that the proposed deep model method outperforms several machine learning algorithms. [ABSTRACT FROM AUTHOR]

Over the past few decades, due to human activities, industrialization, and urbanization, air pollution has become a life-threatening factor in many countries around the world. Among air pollutants, Particulate Matter with a diameter of less than 2.5 μ m ( P M 2.5 ) is a serious health problem. It causes various illnesses such as respiratory tract and cardiovascular diseases. Hence, it is necessary to accurately predict the P M 2.5 concentrations in order to prevent the citizens from the dangerous impact of air pollution beforehand. The variation of P M 2.5 depends on a variety of factors, such as meteorology and the concentration of other pollutants in urban areas. In this paper, we implemented a deep learning solution to predict the hourly forecast of P M 2.5 concentration in Beijing, China, based on CNN-LSTM, with a spatial-temporal feature by combining historical data of pollutants, meteorological data, and P M 2.5 concentration in the adjacent stations. We examined the difference in performances among Deep learning algorithms such as LSTM, Bi-LSTM, GRU, Bi-GRU, CNN, and a hybrid CNN-LSTM model. Experimental results indicate that our method "hybrid CNN-LSTM multivariate" enables more accurate predictions than all the listed traditional models and performs better in predictive performance. [ABSTRACT FROM AUTHOR]

Short-term prediction of origin–destination (OD) flow is a primary but complex assignment to urban rail companies, which is the basis of intelligent and real-time urban rail transit (URT) operation and management. The short-term prediction of URT OD flow has three special characteristics: data lag, data dimensionality, and data malconformation, distinguishing it from other short-term prediction tasks. It is essential to propose a novel prediction algorithm that considers the special characteristics of the URT OD flow. For this purpose, based on deep learning methods and multi-source big data, a modified spatial–temporal long short-term memory (ST-LSTM) model is established. The proposed model comprises four components: (1) a temporal feature extraction module is devised to extract time information within network-wide historical OD data; (2) a spatial correlation learning module is introduced to address the data malconformation and data dimensionality problems, which provides an interpretable spatial correlation quantization method; (3) an input control-gated mechanism is originally proposed to solve the data lag problem, which combines the processed available OD flow and real-time inflow/outflow; (4) a fusion module combines historical spatial–temporal features with real-time information to achieve accurate OD flow prediction. We also further discuss the interpretability of the model in detail. The ST-LSTM model is evaluated by sufficient experiments on two large-scale actual subway datasets from Nanjing and Beijing, and the experimental results demonstrate that it can better learn the spatial–temporal correlations and exceed the rest benchmarking methods. [ABSTRACT FROM AUTHOR]

Road damage is a great threat to the service life and safety of roads, and the early detection of pavement damage can facilitate maintenance and repair. Street view images serve as a new solution for the monitoring of pavement damage due to their wide coverage and regular updates. In this study, a road pavement damage dataset, the Street View Image Dataset for Automated Road Damage Detection (SVRDD), was developed using 8000 street view images acquired from Baidu Maps. Based on these images, over 20,000 damage instances were visually recognized and annotated. These instances were distributed in five administrative districts of Beijing City. Ten well-established object detection algorithms were trained and assessed using the SVRDD dataset. The results have demonstrated the performances of these algorithms in the detection of pavement damages. To the best of our knowledge, SVRDD is the first public dataset based on street view images for pavement damages detection. It can provide reliable data support for future development of deep learning algorithms based on street view images. [ABSTRACT FROM AUTHOR]

Making accurate predictions of subway passenger flow is conducive to optimizing operation plans. This study aims to analyze the regularity of subway passenger flow and combine the modeling skills of deep learning with transportation knowledge to predict the short-term subway passenger flow in the scenarios of workdays and holidays. The processed data were collected from two months of Automated Fare Collection (AFC) data from Xizhimen station of Beijing metro. The data were first cleaned by the established cleansing rules to delete malformed and abnormal logic data. The cleaned data were used to analyze the spatial characteristics in passenger flow. Second, a short-term subway passenger flow prediction model was built on the basis of long short-term memory (LSTM). Determining that the error will be relatively high in peak hours, we proposed gradual optimizations from data input by dividing one whole day into different time periods, and then used particle swarm optimization (PSO) to search for the optimal hyperparameters setting. Finally, inbound passenger flow of Beijing Xizhimen subway station in 2018 was selected for numerical experiments. Predictions of the LSTM-based model had higher accuracy than the traditional machine learning support vector regression (SVR) model, with mean absolute percentage error (MAPE) of 21.97% and 4.80% in the scenarios of workdays and holidays, respectively, which are both lower than those of the SVR model. The optimized PSO-LSTM model has been verified for its effectiveness and accurateness by the AFC data. [ABSTRACT FROM AUTHOR]

Radiotherapy requires the target area and the organs at risk to be contoured on the CT image of the patient. During the process of organs-at-Risk (OAR) of the chest and abdomen, the doctor needs to contour at each CT image. The delineations of large and varied shapes are time-consuming and laborious. This study aims to evaluate the results of two automatic contouring softwares on OARs definition of CT images of lung cancer and rectal cancer patients. The CT images of 15 patients with rectal cancer and 15 patients with lung cancer were selected separately, and the organs at risk were manually contoured by experienced physicians as reference structures. And then the same datasets were automatically contoured based on AiContour (version 3.1.8.0, Manufactured by Linking MED, Beijing, China) and Raystation (version 4.7.5.4, Manufactured by Raysearch, Stockholm, Sweden) respectively. Deep learning auto-segmentations and Atlas were respectively performed with AiContour and Raystation. Overlap index (OI), Dice similarity index (DSC) and Volume difference (Dv) were evaluated based on the auto-contours, and independent-sample t-test analysis is applied to the results. The results of deep learning auto-segmentations on OI and DSC were better than that of Atlas with statistical difference. There was no significant difference in Dv between the results of two software. With deep learning auto-segmentations, auto-contouring results of most organs in the chest and abdomen are good, and with slight modification, it can meet the clinical requirements for planning. With Atlas, auto-contouring results in most OAR is not as good as deep learning auto-segmentations, and only the auto-contouring results of some organs can be used clinically after modification. [ABSTRACT FROM AUTHOR]