Data-Driven Strategy for Merchant Incentive Optimization in Digital Payment Ecosystems

Fully distributed hydrological models take into account the spatial variability of a catchment, allowing for a more accurate representation of its heterogeneity, and assessing its hydrological response at multiple locations. However, physics-based fully distributed models can be time-consuming when it comes to model runtime and calibration, especially for large-scale catchments. On the other hand, deep learning models have shown great potential in the field of hydrological modelling, outperforming lumped rainfall-runoff conceptual models, and improving prediction in ungauged basins via catchment transferability. Despite these advances, the field still lacks a multivariable, fully distributed hydrological deep learning model capable of generalizing to unseen catchments. To address the aforementioned challenges associated with physics-based distributed models and deep learning models, we explore the possibility of developing a fully distributed deep learning model by using Graph Neural Networks (GNN), an extension of deep learning methods to non-Euclidean topologies including graphs and meshes.We develop a surrogate model of wflow_sbm, a fully distributed, physics-based hydrological model, by exploiting the similarities between its underlying functioning and GNNs. The GNN uses the same input as wflow_sbm: distributed static parameters based on physical characteristics of the catchment and gridded dynamic forcings. The GNN is trained to produce the same output as wflow_sbm, predicting multiple gridded variables related to rainfall-runoff, such as streamflow, actual evapotranspiration, subsurface flow, saturated and unsaturated groundwater storage, snow storage, and runoff. We show that our GNN model achieves high performance in unseen catchments, indicating that GNNs are a viable option for fully distributed multivariable hydrological models capable of generalizing to unseen regions. Furthermore, the GNN model achieves a significant computational speedup compared to wflow_sbm. We will continue this research, using the GNN-based surrogate models as pre-trained backbones to be fine-tuned with measured data, ensuring accurate model adaptation, and enhancing their practical applicability in diverse hydrological scenarios.

Machine learning (ML) and deep learning (DL) models can play an important role when it comes to modelling complicated processes. Such capability is necessary for hydrological and climate-related applications. Generally, ML models utilize precipitation and temperature time series of a basin as input to develop a lumped rainfall-runoff model to simulate streamflow at the basin outlet. However, when it is divided into several sub-basins, Graph Neural Networks (GNN) can consider each sub-basin as a node and link them together using a connectivity matrix to account for spatial variations of hydroclimatic variables. In this study, GNN and various ML models with different types of architecture, ranging from neural networks, tree-based structure, and gradient boosting, were exploited for daily streamflow simulation over different case studies. For each case study, the basin was divided into a few sub-basins for which daily precipitation and temperature data were aggregated and used as input. For training GNN, the connection matrix of sub-basins was also used as input. Basically, 75% of historical records were utilized to train GNN and different ML models, e.g., artificial neural networks, support vector machine, decision tree, random forest, eXtreme Gradient Boosting (XGBoost), Light Gradient-Boosting Machine (LightGBM), and Category Boosting (CatBoost), while the rest was used for testing. Streamflow simulation was conducted with/without considering seasonality impact and lag times. The obtained results clearly demonstrate that considering seasonality and time lags can enhance accuracy of streamflow predictions based on Kling–Gupta efficiency (KGE). Furthermore, GNN with seasonality impact and time lags achieved promising results across different case studies with KGE>0.85 for training and KGE>0.59 for testing data, respectively. Among ML models, boosting models, e.g., LightGBM and XGBoost, performed slightly better than other ML models. for Finally, this comparative analysis provides valuable insights for ML/DL applications in climate change impact assessments.Acknowledgements: This research work was carried out as part of the TRANSCEND project with funding received from the European Union Horizon Europe Research and Innovation Programme under Grant Agreement No. 10108411.

This paper conducts an empirical study on detecting faulty sensors in a large-scale sensor network containing approximately 10, 000 sensors distributed over 36, 000 km <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> . First, we discuss the practical challenge of this task. We compare rule-based models, traditional machine learning models, deep learning models without graph neural networks, and deep learning models with graph neural networks. The experimental results show that graph neural networks identify more problematic sensors in fewer trials than rule-based models and other machine learning and deep learning models. In addition to training the models in a central server, we also show that localized versions of the deep learning models with graph neural networks yield predictive power comparable to centralized training. Consequently, each sensor may perform a local inspection to identify its health status and only send reminder signals to a centralized server if it is self-diagnosed as a faulty sensor.

In non-small cell lung cancer (NSCLC), the composition and structure of the tumor immune microenvironment (TME) critically impact patient outcomes, necessitating improved computational approaches to characterize the TME beyond traditional metrics like the PD-L1 tumor proportion score (TPS). Here, we developed an interpretable deep-learning model that robustly predicts patient overall survival (OS) and identifies unique PD-L1 enriched neighborhoods linked to patient prognosis. We utilized a dataset of 507 NSCLC patient biopsy samples prospectively acquired as part of the DFCI ImmunoPROFILE project. Each sample was stained for Cytokeratin as tumor marker, PD-L1, PD-1, CD8, and FOXP3 using a targeted multiplex immunofluorescence assay. Inner tumor regions of interest were automatically processed for spatially resolved identification and quantification of marker expression at the single-cell level. The dataset was divided into 380 patients for model development and 127 for testing. The deep-learning model, a graph neural network (GNN), was trained to predict OS based on local cell neighborhoods using a weakly supervised training paradigm. The GNN models neighborhoods as a graph, with cells and their corresponding marker expression representing nodes and edges connecting adjacent cells, for an average of 42 cells per neighborhood. Survival predictions were obtained by averaging predictions over all neighborhoods from one sample. The GNN model demonstrated robust performance with a concordance index (c-index) of 0.79 on the test data, surpassing traditional metrics such as TPS and immune marker density-based models (c-index: 0.50 and 0.74, respectively). Beyond assessing performance, we investigated the features learned by the GNN using k-means clustering in the model’s feature space. We identified three clusters that were notably enriched in PD-L1 positive tumor cells. The first cluster shows an especially “high TPS” phenotype, where the GNN predicted such neighborhoods to be generally poor for survival. The other two clusters were additionally enriched with PD-L1 positive immune cells (ICs) and associated with favorable survival predictions. Notably, one cluster was also enriched with other types of ICs (”PD-L1 IC mixed”), which the GNN predicted as especially favorable for survival compared to exclusive enrichment of PD-L1 ICs (”PD-L1 IC dominant”). A score summarizing these three PD-L1 expression patterns was significantly associated with OS (p-value: 0.01; HR: 0.72, 95% CI: 0.16-1.27), validating the features learned by the GNN. In sum, we developed a highly interpretable deep-learning model leveraging cellular neighborhood graphs, which effectively identifies expression patterns correlating with NSCLC prognosis, offering promise for developing personalized treatment strategies. Citation Format: Katharina Viktoria Hoebel, James R. Lindsay, Joao V. Alessi, Jason L. Weirather, Ian D. Dryg, Jennifer Altreuter, Mark M. Awad, Scott J. Rodig, William E. Lotter. Deep-learning model trained on multiplex immunofluorescence-stained tissue samples predicts the survival of patients with non-small cell lung cancer better than PD-L1 TPS alone [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 6189.

With the development of science and technology and the improvement of hardware computing power, the application of large models in the field of artificial intelligence (AI) has become a current research hotspot Among the focal points in the field of deep learning, AI for science is one of the highlighted areas, utilizing deep learning methods for pattern recognition, anomaly detection, predictive analysis, and more on a large scale of scientific data. In the realm of materials science, the structure of crystals is composed of edges and nodes, making it readily representable as a graph. In previous research, some typical models, such as the MEGNet model, utilized their graph neural network features to fit computational results based on density functional theory for predicting various material properties. Building on this concept, the authors propose a novel graph neural network (GNN) model, optimized with a Multi-Head Self-Attention (MHSA) mechanism, for predicting materials data with crystal structures. This model is named self-attention enhanced graph neural network. The model segments the input data into three parts: edges, nodes, and global features. The graph convolutional layer module is primarily used for aggregating node, edge, and global features, learning node representations, and capturing higher-order neighborhood information through multiple layers of GNN. The MHSA component allows nodes to learn global dependencies, providing different representation subspaces for the nodes. In comparison with other machine learning and deep learning models, the results indicate an improvement in the predictive accuracy of this model. A new graph neural network (GNN) model called Self-Attention Enhanced Graph Neural Network (SA-GNN) is proposed for predicting the properties of materials with crystal structures. This model incorporates multi-head self-attention to allow nodes to learn global dependencies and generate different representational subspaces. Compared to other machine learning and deep learning models, the results show improved predictive accuracy, demonstrating the potential of graph networks combined with self-attention for modeling crystal material data.

In recent years, deep learning has completely changed many machine learning tasks, and the data in these tasks is usually expressed in Euclidean space. However, as more and more applications need to use non-Euclidean data, vulnerability mining is becoming more and more important. With the successful development of neural networks, many machine learning tasks, such as object detection, image classification, and speech recognition, once relied heavily on manual feature engineering to extract features, and can now be completed with various end-to-end deep learning models, such as Convolutional neural network, long and short-term memory networketc.Vulnerability mining is an important way to prevent and control system vulnerabilities. Traditional methods of vulnerability mining can no longer meet people's needs. In order to enable the vulnerability mining application to meet people's needs, we established a related source code vulnerability mining model based on graph neural networks. By investigating relevant literature, conducting interviews with professionals, etc., collected data from databases such as HowNet, Wanfang Database, SSCI, etc., and built a model of source code vulnerability mining based on graph neural networks using parallel algorithms. Through simulation, we found that the method of mining source code vulnerabilities based on graph neural networks is becoming more and more accepted by people, and the increase in 2016 reached 0.16. Moreover, the efficiency of source code vulnerability mining based on graph neural network is much higher than other vulnerability mining methods, and the mining speed is more than 20% ahead of other mining methods. This shows that source code vulnerability mining based on graph neural network can play an important role in preventing system vulnerabilities.

&amp;lt;p&amp;gt;In the recent past, several studies have demonstrated the ability of deep learning (DL) models, especially based on Long Short-Term Memory (LSTM) networks, for rainfall-runoff modeling. However, almost all of these studies were limited to (multiple) individual catchments or small river networks, consisting of only a few connected catchments.&amp;amp;#160;&amp;lt;/p&amp;gt;&amp;lt;p&amp;gt;In this study, we investigate large-scale, spatially distributed rainfall-runoff modeling using DL models. Our setup consists of two independent model components: One model for the runoff-generation process and one for the routing. The former is an LSTM-based model that predicts the discharge contribution of each sub-catchment in a river network. The latter is a Graph Neural Network (GNN) that routes the water along the river network network in hierarchical order. The first part is set up to simulate unimpaired runoff for every sub-catchment. Then, the GNN routes the water through the river network, incorporating human influences such as river regulations through hydropower plants. The main focus is to investigate different model architectures for the GNN that are able to learn the routing task, as well as potentially accounting for human influence. We consider models based on 1D-convolution, attention modules, as well as state-aware time series models.&amp;lt;/p&amp;gt;&amp;lt;p&amp;gt;The decoupled approach with individual models for sub-catchment discharge prediction and routing has several benefits: a) We have an intermediate output of per-basin discharge contributions that we can inspect. b) We can leverage observed streamflow when available. That is, we can optionally substitute the discharge simulations of the first model with observed discharge, to make use of as much observed information as possible. c) We can train the model very efficiently. d) We can simulate any intermediate node in the river network, without requiring discharge observations.&amp;lt;/p&amp;gt;&amp;lt;p&amp;gt;For the experiments, we use a new large-sample dataset called LamaH (&amp;lt;strong&amp;gt;La&amp;lt;/strong&amp;gt;rge-sa&amp;lt;strong&amp;gt;m&amp;lt;/strong&amp;gt;ple D&amp;lt;strong&amp;gt;a&amp;lt;/strong&amp;gt;ta for &amp;lt;strong&amp;gt;H&amp;lt;/strong&amp;gt;ydrology in Central Europe) that covers all of Austria and the foreign upstream areas of the Danube. We consider the entire Danube catchment upstream of Bratislava, a highly diverse region, including large parts of the Alps, that covers a total area of more than 130000km2. Within that area, LamaH contains hourly and daily discharge observations for more than 600 gauge stations. Thus, we investigate DL-based routing models not only for daily discharge, but also for hourly discharge.&amp;lt;/p&amp;gt;&amp;lt;p&amp;gt;Our first results are promising, both daily and hourly discharge simulation. For example, the fully DL-based distributed models capture the dynamics as well as the timing of the devastating 2002 Danube flood. Building upon our work on learning universal, regional, and local hydrological behaviors with machine learning, we try to make the GNN-based routing as universal as possible, striving towards a globally applicable, spatially distributed, fully learned hydrological model.&amp;lt;/p&amp;gt;

AI-Driven molecule generation and bioactivity prediction: A multi-model approach combining VAE, graph and language-based neural networks.

Precipitation exerts far‐reaching impacts on both socio‐economic fabric and individual well‐being, necessitating concerted efforts in accurate forecasting. Deep learning (DL) models have increasingly demonstrated their prowess in forecasting meteorological elements. However, traditional DL prediction models often grapple with heavy rainfall forecasting. In this study, we propose physics‐informed localized graph neural network models called ω‐GNN and ω‐EGNN, constrained by the coupling of physical variables and climatological background to predict precipitation in China. These models exhibit notable and robust improvements in identifying heavy rainfall while maintaining excellent performance in forecasting light rain by comparing to numerical weather prediction (NWP) and other DL models with multiple perturbation experiments in different data sets. Surprisingly, within a certain range, even when a DL model utilizes more input variables, GNN can still maintain its advantage. The methods to fuse physics into DL model demonstrated in this study may be promising and call for future studies.

The human immunodeficiency virus presents a significant global health challenge due to its rapid mutation and the development of resistance mechanisms against antiretroviral drugs. Recent studies demonstrate the impressive performance of machine learning (ML) and deep learning (DL) models in predicting the drug resistance profile of specific FDA-approved inhibitors. However, generalizing ML and DL models to learn not only from isolates but also from inhibitor representations remains challenging for HIV-1 infection. We propose a novel drug-isolate-fold change (DIF) model framework that aims to predict drug resistance score directly from the protein sequence and inhibitor representation. Various ML and DL models, inhibitor representations, and protein representations were analyzed through realistic validation mechanisms. To enhance the molecular learning capacity of DIF models, we employ a transfer learning approach by pretraining a graph neural network (GNN) model for activity prediction on a data set of 4855 HIV-1 protease inhibitors (PIs). By performing various realistic validation strategies on internal and external genotype-phenotype data sets, we statistically show that the learned representations of inhibitors improve the predictive ability of DIF-based ML and DL models. We achieved an accuracy of 0.802, AUROC of 0.874, and r of 0.727 for the unseen external PIs. By comparing the DIF-based models with a null model consisting of isolate-fold change (IF) architecture, it is observed that the DIF models significantly benefit from molecular representations. Combined results from various testing strategies and statistical tests confirm the effectiveness of DIF models in testing novel PIs for drug resistance in the presence of an isolate.

Graph Neural Networks (GNNs) are deep learning models that take graph data as inputs, and they are applied to various tasks such as traffic prediction and molecular property prediction. However, owing to the complexity of the GNNs, it has been difficult to analyze which parts of inputs affect the GNN model’s outputs. In this study, we extend explainability methods for Convolutional Neural Networks (CNNs), such as Local Interpretable Model-Agnostic Explanations (LIME), Gradient-Based Saliency Maps, and Gradient-Weighted Class Activation Mapping (Grad-CAM) to GNNs, and predict which edges in the input graphs are important for GNN decisions. The experimental results indicate that the LIME-based approach is the most efficient explainability method for multiple tasks in the real-world situation, outperforming even the state-of-the-art method in GNN explainability.

In this work, we propose a novel motor disorder diagnosis model based on graph neural networks (GNNs). This model maximizes model performance by incorporating advanced preprocessing techniques such as Fast Fourier Transform (FFT) and Wavelet Transform (WT). Conventional machine learning and deep learning models such as CNN and SVM find it difficult to handle nonlinear high-dimensional data in motor disorder diagnosis. On the other hand, GNN effectively handles these complex data structures, enabling more accurate and reliable defect classification. Experimental results show that the GNN-based model combining FFT and WT performed well in the diagnosis of motor disorder. Specifically, the FFT-based GNN showed high accuracy, accuracy, and reproducibility at an F1 score of 0.95. The GNN model has lower misclassification rate and higher reliability compared to other models, and ran consistently for various defect types. This is because GNNs can capture complex relationships within frequency domain function (FFT) and time frequency domain pattern (WT). For example, rotational imbalance defects are accurately classified thanks to the ability of GNNs to model harmonic frequency relationships, and bearing defects are accurately classified thanks to the model sensitivity to local frequency spikes that are effectively represented on nodes and edges of the graph. These results suggest that GNN-based motor defect diagnostic systems not only improve diagnostic accuracy, but also have significant potential for real-time applications in manufacturing environments. The system is expected to reduce maintenance costs and improve operational efficiency. The proposed GNN model makes an important contribution by providing practical solutions for the detection and prevention of motion disorders.

Developing region-specific radar quantitative precipitation estimation (QPE) products for South China (SC) is crucial&amp;#160;due to its unique climate and complex terrain over there. Deep learning (DL) has emerged as a promising avenue for radar QPE, especially graph neural networks (GNNs). Many studies have tested the DL models in radar QPE, but virtually no studies have evaluated the performance of DL models in different precipitation intensities, types, or organizations. Moreover, limited attention has been given to whether DL-based methods can mitigate radar QPE errors caused by orographic influences in complex terrains, such as those in SC.This study investigates the advantages of DL methods for QPE tasks in South China, utilizing nearly three years of hourly gauge data as labels and ground-based radar reflectivity as inputs. Firstly, multi-layer perceptron (MLP), Convolutional Neural Networks (CNNs), and GNNs with similar architectures are constructed and compared to traditional Z-R relationships considering precipitation types. DL methods outperform traditional Z-R relationships and GNNs perform the best. More importantly, this study conducts a systematic evaluation of the proposed GNN. For extreme precipitation (&gt;30 mm/h), GNN achieves the smallest MAE, highlighting its potential for hazardous event estimation. It also demonstrates stable performance for stratiform and organized precipitation, with minimal bias and standard deviation. However, GNN is less effective for isolated precipitation, whereas CNNs are a better choice due to their ability to estimate scattered rainfall accurately. Last but not least, the Z-R relationship shows systematic spatial biases, overestimating precipitation in coastal plains and underestimating it in inland high-altitude regions. DL methods alleviate these terrain-induced biases by incorporating spatial information. Overall, this study highlights the advantages of DL methods across different precipitation scenarios and demonstrates their ability to mitigate systematic biases from complex terrain.

The advance of deep learning has shown great potential in applications (speech, image, and video classification). In these applications, deep learning models are trained by datasets in Euclidean space with fixed dimensions and sequences. Nonetheless, the rapidly increasing demands on analyzing datasets in non‐Euclidean space require additional research. Generally speaking, finding the relationships of elements in datasets and representing such relationships as weighted graphs consisting of vertices and edges is a viable way of analyzing datasets in non‐Euclidean space. However, analyzing the weighted graph‐based dataset is a challenging problem in existing deep learning models. To address this issue, graph neural networks (GNNs) leverage spectral and spatial strategies to extend and implement convolution operations in non‐Euclidean space. Based on graph theory, a number of enhanced GNNs are proposed to deal with non‐Euclidean datasets. In this study, we first review the artificial neural networks and GNNs. We then present ways to extend deep learning models to deal with datasets in non‐Euclidean space and introduce the GNN‐based approaches based on spectral and spatial strategies. Furthermore, we discuss some typical Internet of Things (IoT) applications that employ spectral and spatial convolution strategies, followed by the limitations of GNNs in the current stage.

Deep learning methods have been successfully applied to the tasks of predicting functional genomic elements such as histone marks, transcriptions factor binding sites, non-B DNA structures, and regulatory variants. Initially convolutional neural networks (CNN) and recurrent neural networks (RNN) or hybrid CNN-RNN models appeared to be the methods of choice for genomic studies. With the advance of machine learning algorithms other deep learning architectures started to outperform CNN and RNN in various applications. Graph neural network (GNN) applications improved the prediction of drug effects, disease associations, protein-protein interactions, protein structures and their functions. The performance of GNN is yet to be fully explored in genomics. Earlier we developed DeepZ approach in which deep learning model is trained on information both from sequence and omics data. Initially this approach was implemented with CNN and RNN but is not limited to these classes of neural networks. In this study we implemented the DeepZ approach by substituting RNN with GNN. We tested three different GNN architectures - Graph Convolutional Network (GCN), Graph Attention Network (GAT) and inductive representation learning network GraphSAGE. The GNN models outperformed current state-of the art RNN model from initial DeepZ realization. Graph SAGE showed the best performance for the small training set of human Z-DNA ChIP-seq data while Graph Convolutional Network was superior for specific curaxin-induced mouse Z-DNA data that was recently reported. Our results show the potential of GNN applications for the task of predicting genomic functional elements based on DNA sequence and omics data.Availability and implementation–The code is freely available at https://github.com/MrARVO/GraphZ.