Representation Of Interactions Research Articles

Improvements in the land surface configuration to better simulate seasonal snow cover in the European Alps with the CNRM-AROME (cycle 46) convection-permitting regional climate model

Abstract. Snow cover modeling remains a major challenge in climate and numerical weather prediction (NWP) models even in recent versions of high-resolution coupled surface–atmosphere (i.e., at kilometer scale) regional models. Evaluation of recent climate simulations, carried out as part of the WCRP-CORDEX Flagship Pilot Study on Convection (FPSCONV) with the CNRM-AROME convection-permitting regional climate model at 2.5 km horizontal resolution, has highlighted significant snow cover biases, severely limiting its potential in mountain regions. These biases, which are also found in AROME numerical weather prediction (NWP) model results, have multiple causes, involving atmospheric processes and their influence on input data to the land surface models in addition to deficiencies of the land surface model itself. Here we present improved configurations of the SURFEX-ISBA land surface model used in CNRM-AROME. We thoroughly evaluated these configurations on their ability to represent seasonal snow cover across the European Alps. Our evaluation was based on coupled simulations spanning the winters of 2018–2019 and 2019–2020, which were compared against remote sensing data and in situ observations. More specifically, the study tests the influence of various changes in the land surface configuration, such as the use of multi-layer soil and snow schemes, the division of the energy balance calculation by surface type within a grid cell (multiple patches), and new physiographic databases and parameter adjustments. Our findings indicate that using only more detailed individual components in the surface model did not improve the representation of snow cover due to limitations in the approach used to account for partial snow cover within a grid cell. These limitations are addressed in further configurations that highlight the importance, even at kilometer resolution, of taking into account the main subgrid surface heterogeneities and improving representations of interactions between fractional snow cover and vegetation. Ultimately, we introduce a land surface configuration, which substantially improves the representation of seasonal snow cover in the European Alps in coupled CNRM-AROME simulations. This holds promising potential for the use of such model configurations in climate simulations and numerical weather prediction both for AROME and other high-resolution climate models.

Multimodal meaning making in news communication about immigration: using the NewsScape corpus to explore co-verbal images in TV news

The communication of news relies on semiotic resources besides language, including various audiovisual modes of representation. Owing to the difficulties associated with obtaining televisual data, the vast majority of research addressing multimodality in the news has been targeted at print news media, where various strategies in visual representation and patterns of interaction between verbal and visual modes have been discerned. Where televisual data has been interrogated, this has been based on a very limited number of data points. In this study, I exploit the NewsScape library – a massive multimodal corpus of news communication – to investigate multimodal representations of immigration in television news. Accessible via CQPWeb, the corpus is searched for target utterances refugees/(im)migrants have VERBed and refugees/(im)migrants are VERBing. The co-verbal images accompanying 474 utterances describing motion events are analyzed quantitatively and qualitatively. Among the results discussed are that refugees/migrants are depicted in large rather than small groups, that they are depicted in transit somewhere along the migratory journey rather than in countries of origin or destination countries, that they are depicted on land more than at sea, that they are depicted in security contexts, and that they are erased represented instead through abstract forms such as maps. Differences in the visual representation of people designated as ‘refugee’ versus ‘migrant’ are also observed and discussed.

Tripartite interaction representation algorithm for crystal graph neural networks

Driven by emerging research paradigms, the application of artificial intelligence models presents innovative tools for designing materials and optimizing their performance. In the field of materials science, there is a current research emphasis on exploring techniques for characterizing material structures to achieve precise descriptions. This paper proposes a crystal graph convolution neural network model that incorporates a tripartite interaction approach. The model not only incorporates atomic information, bond lengths, and bond angles but also offers a method for updating atoms and bond lengths, facilitating accurate descriptions of crystal structures by capturing implicit structural information. Focusing on predicting the formation energy of crystalline compounds, our results demonstrate improved predictive accuracy compared to existing representation algorithms. The average error of the formation energy in the random dataset, demonstrating robust generalization, is merely 0.048eV/atom, with an impressive R2\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$R^{2}$$\\end{document} value of 0.994. Additionally, this paper establishes a crystal graph neural network framework for predicting algorithm performance. By integrating automatic parallel algorithms and an automated process, we achieve a synthesis of these techniques, enhancing computational efficiency and streamlining algorithm usage.

Resource sharing of an infant gut microbiota synthetic community in combinations of human milk oligosaccharides.

Quickly after birth, the gut microbiota is shaped via species acquisition and resource pressure. Breastmilk, and more specifically, human milk oligosaccharides are a determining factor in the formation of microbial communities and the interactions between bacteria. Prominent human milk oligosaccharide degraders have been rigorously characterized, but it is not known how the gut microbiota is shaped as a complex community. Here, we designed BIG-Syc, a synthetic community of 13 strains from the gut of vaginally born, breastfed infants. BIG-Syc replicated key compositional, metabolic, and proteomic characteristics of the gut microbiota of infants. Upon fermentation of a 4 and 5 human milk oligosaccharide mix, BIG-Syc demonstrated different compositional and proteomic profiles, with Bifidobacterium infantis and Bifidobacterium bifidum suppressing one another. The mix of 5 human milk oligosaccharides resulted in a more diverse composition with dominance of B. bifidum, whereas that with 4 human milk oligosaccharides supported the dominance of B. infantis, in 4 of 6 replicates. Reintroduction of bifidobacteria to BIG-Syc led to their engraftment and establishment of their niche. Based on proteomics and genome-scale metabolic models, we reconstructed the carbon source utilization and metabolite and gas production per strain. BIG-Syc demonstrated teamwork as cross-feeders utilized simpler carbohydrates, organic acids, and gases released from human milk oligosaccharide degraders. Collectively, our results showed that human milk oligosaccharides prompt resource-sharing for their complete degradation while leading to a different compositional and functional profile in the community. At the same time, BIG-Syc proved to be an accurate model for the representation of intra-microbe interactions.

OriTDB: a database of the origin-of-transfer regions of bacterial mobile genetic elements.

Conjugation and mobilization are two important pathways of horizontal transfer of bacterial mobile genetic elements (MGEs). The origin-of-transfer (oriT) region is crucial for this process, serving as a recognition site for relaxase and containing the DNA nicking site (nic site), which initiates the conjugation or mobilization. Here, we present a database of the origin-of-transfer regions of bacterial MGEs, oriTDB (https://bioinfo-mml.sjtu.edu.cn/oriTDB2/). Incorporating data from text mining and genome analysis, oriTDB comprises 122 experimentally validated and 22927 predicted oriTs within bacterial plasmids, Integrative and Conjugative Elements, and Integrative and Mobilizable Elements. Additionally, oriTDB includes details about associated relaxases, auxiliary proteins, type IV coupling proteins, and a gene cluster encoding the type IV secretion system. The database also provides predicted secondary structures of oriT sequences, dissects oriT regions into pairs of inverted repeats, nic sites, and their flanking conserved sequences, and offers an interactive visual representation. Furthermore, oriTDB includes an enhanced oriT prediction pipeline, oriTfinder2, which integrates a functional annotation module for cargo genes in bacterial MGEs. This resource is intended to support research on bacterial conjugative or mobilizable elements and promote an understanding of their cargo gene functions.

Investigation of ultrasonic propagation in high attenuation porous materials

This research examines ultrasonic wave propagation in air-saturated plastic foams used for noise pollution reduction, questioning the effectiveness of current models such as the well known Johnson-Champoux-Allard model at high frequencies. These foams present a challenge for existing models to accurately depict visco-inertial and thermal interactions within their pores. The study highlights the model's failure to align experimental results with theoretical predictions. It introduces novel parameters denoted as Σ and V for viscous effects, and Σ′ and V ′ for thermal effects, to improve the representation of fluid-sturcture interactions. These parameters suggest a more significant boundary layer effect within the pores than previously considered. This approach aims to provide additional physical context for the principles governing the behavior of highly attenuating porous media and to explore new avenues for material characterization that surpass the limitations of existing models.

Interactive external representations for knowledge-building processes and performance in dyads: An experimental study

Interactive external representations for knowledge-building processes and performance in dyads: An experimental study

Bi-Branching Feature Interaction Representation Learning for Multivariate Time Series

Bi-Branching Feature Interaction Representation Learning for Multivariate Time Series

When latent features meet side information: A preference relation based graph neural network for collaborative filtering

As recommender systems shift from rating-based to interaction-based models, graph neural network-based collaborative filtering models are gaining popularity due to their powerful representation of user-item interactions. However, these models may not produce good item ranking since they focus on explicit preference predictions. Further, these models do not consider side information since they only capture latent feature information of user-item interactions. This study proposes an approach to overcome these two issues by employing preference relation in the graph neural network model for collaborative filtering. Using preference relation ensures the model will generate a good ranking of items. The item side information is integrated into the model through a trainable matrix, which is crucial when the data is highly sparse. The main advantage of this approach is that the model can be generalized to any recommendation scenario where a graph neural network is used for collaborative filtering. Experimental results obtained using the recent RS datasets show that the proposed model outperformed the related baselines.

How to Help Older Adults Navigate Through VR? Effects in Outdoor–Indoor Transition Environments

In more complex outdoor–indoor transition scenarios, people were more prone to getting lost due to difficulties in effectively integrating proprioceptive cues and external navigation cues. Previous research had mostly focused on spatial perception driven by head or limb movements, with less discussion on locomotion methods that involved significant visual discrepancies. The challenge in designing navigation systems lies in balancing the enhancement of wayfinding performance with the potential negative impact on spatial cognition. We believe that navigation legibility might be a key optimization approach. Additionally, the differences in performance between different age groups were also worth investigating. Therefore, this study aimed to explore how locomotion methods and navigation legibility affected wayfinding performance, navigation interaction, navigation regression, and spatial cognition in younger and older adults during outdoor–indoor transition scenarios. Twenty-four younger (aged 19–27) and 24 older adults (aged 60–83) underwent VR wayfinding tasks. Participants initially followed virtual phone navigation from outdoor starting points to indoor endpoints, then drew cognitive route maps. Five main findings emerged: Continuous teleportation positively affected navigation performance and global representation of routes. Improving navigation legibility boosted navigation interaction, navigation regression, and global representation of routes. Older adults engaged more in navigation interaction but less in navigation regression, exhibiting poorer performance and route representation. Location-specific differences were observed in locomotion, legibility, and age effects. Furthermore, most older adults and over half of younger adults displayed a deflection effect based on outdoor–indoor (OI) space in route representation. These findings indicated that continuous teleportation and enhanced navigation legibility could improve user experience and spatial cognition, highlighting the importance of considering user age and the characteristics of outdoor–indoor transition scenarios when optimizing navigation methods.

Abstract B093: The evaluation of the gene expression profile of pancreatic ductal adenocarcinomas (PDACs) in 3D cell culture vs in vivo and development of new 3D PDACs' models for an evaluation of anti-cancer drug testing

Abstract In the ever-evolving landscape of biomedical research, 3D Cell Culture (3DCC) systems have emerged as pivotal tools for mimicking in vivo environments and fostering more physiologically relevant cellular responses. Despite promises, 3DCC based drug discovery efforts didn’t fulfil the hopes. The classic way 3DCC is done, always starts with 2D cell culture (2DCC) followed by brief 3 day 3DCC and finally drug testing. Following those observations we decided to mimic much more than just brief 3DCC state but enhance the culture by constant 3DCC to 3DCC passages. We characterized the gene expression profiles and proteomic characteristics of pancreatic tumors concluding that prolonged 3DCC is the closest to an in vivo model. The protracted cultivation periods mimic the intricacies of the in vivo environment, enabling a more accurate representation of the molecular dynamics within pancreatic tumors. In our study, we identified a panel of genes characterized by similar expression levels between prolonged 3D cell culture and in vivo conditions, while displaying significant divergence between 2D cultures and short-term 3D cultures when compared to in vivo. These genes play a pivotal role in the progression and treatment of cancer, underscoring the potential of prolonged 3DCC of LifeGel® significance in understanding tumor biology and devising therapeutic strategies. Furthermore, we also investigated the implications of prolonged 3D in vitro cultures on evaluating anti-cancer drug activities. The findings highlight the significance of 3DCC in enhancing our understanding of pancreatic tumor biology and in refining drug screening methodologies for more effective therapeutic interventions. By providing a biomimetic microenvironment, our hydrogels facilitate the formation of complex cellular structures, enabling more accurate representation of tissue architecture and cellular interactions. We delivered LifeGel® platform which presents a myriad of opportunities for researchers seeking to enhance the fidelity and translatability of their in vitro studies into pre-clinial models. Citation Format: Marcin Krzykawski, David Earnshaw, Krzysztof Kus, Artur Pirog, Sachin Kote, Mark Treherne, Justyna Kocik-Krol. The evaluation of the gene expression profile of pancreatic ductal adenocarcinomas (PDACs) in 3D cell culture vs in vivo and development of new 3D PDACs' models for an evaluation of anti-cancer drug testing [abstract]. In: Proceedings of the AACR Special Conference in Cancer Research: Advances in Pancreatic Cancer Research; 2024 Sep 15-18; Boston, MA. Philadelphia (PA): AACR; Cancer Res 2024;84(17 Suppl_2):Abstract nr B093.

Graph contrast learning for recommendation based on relational graph convolutional neural network

Current knowledge graph-based recommendation methods heavily rely on high-quality knowledge graphs, often falling short in effectively addressing issues such as the cold start problem and heterogeneous noise in user interactions. This leads to biases in user interest and popularity. To overcome these challenges, this paper introduces a novel recommendation approach termed Knowledge-enhanced Perceptive Graph Attention with Graph Contrastive Learning (KPA-GCL), which leverages relational graph convolutional neural networks. The proposed method optimizes the triplet embedding representation of entity-item interactions based on relationships between adjacent entities in a heterogeneous graph. Subsequently, a graph convolutional neural network is employed for enhanced aggregation. Similarity scores from a contrastive view serve as the selection criterion for high-quality embedded representations, facilitating the extraction of refined knowledge subgraphs. Multiple adaptive contrast-loss optimization functions are introduced by combining Bayesian Personalized Ranking (BPR) and hard negative sampling techniques. Comparative experiments are conducted with ten popular existing methods using real public datasets. Results indicate that the KPA-GCL method outperforms compared methods in all datasets based on Recall, NDCG, Precision, and Hit-ratio measures. Furthermore, in terms of mitigating cold start and noise, the KPA-GCL method surpasses other ten methods. This validates the reasonability and effectiveness of KPA-GCL in real-world datasets.

“They are here. They are everywhere. They are us.” – Posthuman Encounters in Samanta Schweblin’s Little Eyes (2018)

In Little Eyes (2018), the latest technological hype comes in the shape of cute, pet-like robots with cameras for eyes. These so-called kentukis are remotely inhabited and controlled by their human users via an online connection which is established at random. As the novel’s blurb - “They are here. They are everywhere. They are us.” - suggests, a kentuki is at once a familiar and unfamiliar creature and users’ experiences range from comforting to unsettling. The novel revolves around the theme of stranger danger and reports several uncanny encounters between humans and not-quite-humans in places around the world. The representation of these posthuman interactions in the novel remains ambiguous: Even though the potential to challenge or even transgress the human/non-human binary is addressed, the novel follows a classic dystopian narrative, which posits that it is not the technology itself which is inherently good or bad, it is us humans.

Comparing the interactions between particulate matter and cloud properties over two populated cities in Texas using WRF-Chem fine-resolution modeling

Accurate modeling of aerosol-cloud interactions is essential for reliable weather and air quality simulations, given their significant impact on precipitation patterns, cloud dynamics, and aerosol distributions. This study employed the Weather Research and Forecasting model coupled with Chemistry (WRF-Chem) to examine the impact of enhanced meteorological simulations, achieved through advanced microphysics parameterization supported by data assimilation techniques, on air quality across Texas on August 19 and 20, 2022. We tested four distinct configurations: (1) the Morrison two-moment bulk microphysics scheme, (2) Morrison's with observation nudging, (3) the Spectral Bin Microphysics (SBM), and (4) SBM with observation nudging. While the SBM scheme is known for its detailed representation of aerosol-cloud interactions, our focus was on how improvements in meteorological accuracy translate to more precise air quality simulations. Our findings demonstrated a progressive improvement in simulation accuracy, starting with the Morrison's scheme and further enhanced by adopting the SBM scheme, complemented by incorporating observation nudging. Specifically, the combination of the SBM scheme and the nudging substantially enhanced the model's ability to capture convective precipitation events, as shown by better alignment with NEXRAD radar reflectivity, with R increasing from −0.21 to 0.82, IOA from 0.10 to 0.87, and NMB decreasing from 99% to 34% in Houston. The enhanced meteorology translated into more accurate PM2.5 concentration simulations, particularly through the more accurate representation of aerosol washout during precipitation events. In Houston, the SBM scheme with nudging improved the model's PM2.5 simulations significantly, with NMB decreasing from −20% to 5% and IOA improving from 0.43 to 0.61. In San Antonio, improvements were also notable, with NMB improved from −27% to −22%, R increased from 0.48 to 0.82, and IOA increased from 0.66 to 0.86. Our results underscore the crucial role of accurate meteorological simulations in refining our understanding of aerosol behaviors in relation to precipitation patterns, directly enhancing the reliability and effectiveness of air quality modeling.

On the contribution of local anisotropic creep to macroscopic irradiation-induced growth in zirconium alloy polycrystals

ABSTRACT Zirconium alloys used in nuclear reactor exhibit, under fast neutron flux, a macroscopic deformation even without applied stress called irradiation induced growth. Because of the polycrystalline nature of the material, the local growth of individual grains results in strain incompatibilities yielding to intergranular stresses and thus to local creep. In order to study and understand the influence of the local anisotropic creep on the macroscopic growth behaviour, an analytical and a numerical study have been undertaken, using Voigt and self-consistent estimates and also fast Fourier transform simulations. It is shown that the anisotropic local creep has a strong influence on the effective macroscopic growth strain of the polycrystal. Especially, when the deformation is difficult along the 〈 c 〉 axis, a growth enhancement effect is observed. This phenomenon is well explained in the frame of the Voigt estimate using a reduced fibre texture. Computations conducted using a texture representative of the industrial material provide a quantitative confirmation of this enhancement effect. This work demonstrates the significant contribution of the local anisotropic creep to the macroscopic in-reactor growth strain of zirconium alloys. Highlights The local creep anisotropy has a strong influence on macroscopic irradiation growth of polycrystalline zirconium tubes. The macroscopic axial growth can be significantly enhanced by the interactions between grains when local creep deformation is prevented along the 〈 c 〉 axis. A simplified representation of grain interactions (reduced fibre texture and Voigt estimate) offers a qualitative understanding of this phenomenon. Refined estimates of the effective growth strain are provided for realistic crystallographic texture using polycrystalline unit-cell FFT computations.

Enhancing the synergy governance capability of urban communities based on Bayesian network: A case of Jinan city

The rapid pace of urbanization has brought about a plethora of challenges for urban communities, prompting the advocacy of collaborative governance as a means to tackle management issues effectively. However, fostering strong collaboration among multiple governance entities remains a significant hurdle in urban communities governance. This study addresses this challenge by exploring the complex structure of urban communities' collaborative governance systems, with a focus on enhancing collaborative governance capability. Drawing on Bayesian network (BN) theory and utilizing the case of Jinan, this article investigates the intrinsic mechanisms of achieving collaboration in urban communities. This study establishes a Bayesian network model for collaborative governance of urban communities, providing a visual representation of the intricate interactions among multiple governance entities. We employ a data-knowledge-driven approach for BN structure learning. This approach tackles issues such as the lack of cyclic dependency capability and the impact of small datasets on model accuracy. Furthermore, we analyze the collaborative mechanism through Bayesian reasoning. The findings underscore the pivotal entities and five key factors that significantly influence collaborative governance capability. Lastly, policy implications are drawn to enhance the collaborative governance capability of urban communities.

MGNDTI: A Drug-Target Interaction Prediction Framework Based on Multimodal Representation Learning and the Gating Mechanism.

Drug-Target Interaction (DTI) prediction facilitates acceleration of drug discovery and promotes drug repositioning. Most existing deep learning-based DTI prediction methods can better extract discriminative features for drugs and proteins, but they rarely consider multimodal features of drugs. Moreover, learning the interaction representations between drugs and targets needs further exploration. Here, we proposed a simple M ulti-modal G ating N etwork for DTI prediction, MGNDTI, based on multimodal representation learning and the gating mechanism. MGNDTI first learns the sequence representations of drugs and targets using different retentive networks. Next, it extracts molecular graph features of drugs through a graph convolutional network. Subsequently, it devises a multimodal gating network to obtain the joint representations of drugs and targets. Finally, it builds a fully connected network for computing the interaction probability. MGNDTI was benchmarked against seven state-of-the-art DTI prediction models (CPI-GNN, TransformerCPI, MolTrans, BACPI, CPGL, GIFDTI, and FOTF-CPI) using four data sets (i.e., Human, C. elegans, BioSNAP, and BindingDB) under four different experimental settings. Through evaluation with AUROC, AUPRC, accuracy, F1 score, and MCC, MGNDTI significantly outperformed the above seven methods. MGNDTI is a powerful tool for DTI prediction, showcasing its superior robustness and generalization ability on diverse data sets and different experimental settings. It is freely available at https://github.com/plhhnu/MGNDTI.

Impacts of forest canopy heterogeneity on plot-scale hydrometeorological variables - Insights from an experiment in the humid boreal forest with the Canadian Land Surface Scheme

High latitude regions, including the circumpolar boreal biome, are experiencing important changes in the availability of usable surface water because of climate change. In this context, an adequate representation of the land-atmosphere interaction is critical to ensure optimal management of current and future water resources, forest management, and climate prediction. However, the task is particularly intricate in high-latitude boreal forest, as land surface model faces several challenges due to the unique environmental conditions and ecological characteristics. The objective of this study is to quantify the impact of forest landscape heterogeneity, specifically stand leaf-area index (LAI), soil texture, and drainage regime, on surface water and energy balance in a small boreal high-latitude sub-catchment. To this end, hydrometeorological conditions at seventeen 20×20 m plots in a 1-km2 boreal forest sub-basin are simulated using the Canadian Land Surface Scheme (CLASS), a land surface model, at the point scale. The subplot-scale soil texture, drainage regime, and vegetation characteristics and type are based as closely as possible on field measurements and observations for the 17 plots. The model-driven experiment comprises two sets of simulations using CLASS, each employing the same model setup and run for the 17 experimental plots. The main set employs meteorological forcing from a local micrometeorological tower within the sub-basin to investigate the plot-to-plot variability of albedo, energy fluxes, and soil state variables. A second set of simulations is conducted using meteorological forcing from the ERA5-Land reanalysis, which spans from 1986 to 2022. This data provides a longer time series, enabling a more accurate representation of the interannual climatic variability in the sub-basin. The results of the main and secondary sets of CLASS simulations are used to assess the plot-to-plot and temporal variability of several key hydrometeorological variables by calculating a monthly spread. In brief, the following conclusions and broader implications can be drawn from the findings: i) The simulated total annual evapotranspiration remains relatively uniform between plots despite notable variation in its partitioning from plot to plot. ii) In the presence of a full snowpack, the albedo exhibits substantial heterogeneity at the subplot scale, linked to the canopy's LAI. iii) Local soil properties, drainage regime, and vegetation structure and type exhibit substantial influence on the plot-to-plot variability in soil water content. iv) When parameterized with localized observations and measurements, CLASS can represent and be responsive to the complex dynamics of energy and water fluxes at the plot scale within the heterogeneous surface of boreal forests.

Stargazing Live!: Inspiring with semi-live astronomy data; teaching curriculum topics using smart education tools

Stargazing Live! aims to capture the imagination of learners with a combination of live and interactive planetarium lessons, real astronomical data, and lessons built around interactive knowledge representations. The lessons were created using a co-creation model and tackle concepts in the pre-university (astro)physics which students find difficult to grasp with traditional interventions. An evaluation study in 9 Dutch classrooms showed that learners are inspired and engaged by the planetarium lessons but are not always able to link the content to the classroom. Pre- and post-tests showed that the accompanying star properties activity significantly increased learners’ understanding of the causal relationships between mass and other properties (such as luminosity, gravity, and temperature) in a main sequence star.

Graph Neural Networks for Predicting Solubility in Diverse Solvents Using MolMerger Incorporating Solute-Solvent Interactions.

The prediction of solubility is a complex and challenging physicochemical problem that has tremendous implications for the chemical and pharmaceutical industry. Recent advancements in machine learning methods have provided a great scope for predicting the reliable solubility of a large number of molecular systems. However, most of these methods rely on using physical properties obtained from experiments and expensive quantum chemical calculations. Here, we developed a method that utilizes a graphical representation of solute-solvent interactions using "MolMerger," which captures the strongest polar interactions between molecules using Gasteiger charges and creates a graph incorporating the true nature of the system. Using these graphs as input, a neural network learns the correlation between the structural properties of a molecule in the form of node embedding and its physicochemical properties as the output. This approach has been used to calculate molecular solubility by predicting the Log solubility values of various organic molecules and pharmaceuticals in diverse sets of solvents.