Predicting drug-target binding affinity with cross-scale graph contrastive learning.

Identifying the binding affinity between a drug and its target is essential in drug discovery and repurposing. Numerous computational approaches have been proposed for understanding these interactions. However, most existing methods only utilize either the molecular structure information of drugs and targets or the interaction information of drug-target bipartite networks. They may fail to combine the molecule-scale and network-scale features to obtain high-quality representations. In this study, we propose CSCo-DTA, a novel cross-scale graph contrastive learning approach for drug-target binding affinity prediction. The proposed model combines features learned from the molecular scale and the network scale to capture information from both local and global perspectives. We conducted experiments on two benchmark datasets, and the proposed model outperformed existing state-of-art methods. The ablation experiment demonstrated the significance and efficacy of multi-scale features and cross-scale contrastive learning modules in improving the prediction performance. Moreover, we applied the CSCo-DTA to predict the novel potential targets for Erlotinib and validated the predicted targets with the molecular docking analysis.

CATM: A Multi-Feature-Based Cross-Scale Attentional Convolutional EEG Emotion Recognition Model.

Aiming at the problem that existing emotion recognition methods fail to make full use of the information in the time, frequency, and spatial domains in the EEG signals, which leads to the low accuracy of EEG emotion classification, this paper proposes a multi-feature, multi-frequency band-based cross-scale attention convolutional model (CATM). The model is mainly composed of a cross-scale attention module, a frequency-space attention module, a feature transition module, a temporal feature extraction module, and a depth classification module. First, the cross-scale attentional convolution module extracts spatial features at different scales for the preprocessed EEG signals; then, the frequency-space attention module assigns higher weights to important channels and spatial locations; next, the temporal feature extraction module extracts temporal features of the EEG signals; and, finally, the depth classification module categorizes the EEG signals into emotions. We evaluated the proposed method on the DEAP dataset with accuracies of 99.70% and 99.74% in the valence and arousal binary classification experiments, respectively; the accuracy in the valence-arousal four-classification experiment was 97.27%. In addition, considering the application of fewer channels, we also conducted 5-channel experiments, and the binary classification accuracies of valence and arousal were 97.96% and 98.11%, respectively. The valence-arousal four-classification accuracy was 92.86%. The experimental results show that the method proposed in this paper exhibits better results compared to other recent methods, and also achieves better results in few-channel experiments.

CSINet: A Cross-Scale Interaction Network for Lightweight Image Super-Resolution.

In recent years, advancements in deep Convolutional Neural Networks (CNNs) have brought about a paradigm shift in the realm of image super-resolution (SR). While augmenting the depth and breadth of CNNs can indeed enhance network performance, it often comes at the expense of heightened computational demands and greater memory usage, which can restrict practical deployment. To mitigate this challenge, we have incorporated a technique called factorized convolution and introduced the efficient Cross-Scale Interaction Block (CSIB). CSIB employs a dual-branch structure, with one branch extracting local features and the other capturing global features. Interaction operations take place in the middle of this dual-branch structure, facilitating the integration of cross-scale contextual information. To further refine the aggregated contextual information, we designed an Efficient Large Kernel Attention (ELKA) using large convolutional kernels and a gating mechanism. By stacking CSIBs, we have created a lightweight cross-scale interaction network for image super-resolution named "CSINet". This innovative approach significantly reduces computational costs while maintaining performance, providing an efficient solution for practical applications. The experimental results convincingly demonstrate that our CSINet surpasses the majority of the state-of-the-art lightweight super-resolution techniques used on widely recognized benchmark datasets. Moreover, our smaller model, CSINet-S, shows an excellent performance record on lightweight super-resolution benchmarks with extremely low parameters and Multi-Adds (e.g., 33.82 dB@Set14 × 2 with only 248 K parameters).

Differentiated impacts of environmental contexts on residents' environmental attitudes towards ecological restoration programs of China's drylands.

Residents' environmental attitudes (EAs) towards ecological restoration programs are vital for evaluating program effectiveness and promoting environmental management. However, most local studies have neglected the indirect environmental contextual influences on residents' EAs, and have omitted the regional variations in the environmental contextual influences. To investigate the multilevel factors affecting residents' EAs, we conducted a transect survey that included the eastern, middle, and western regions in northern China's drylands, where have experienced ecological restoration. Multilevel linear models (MLMs) were applied to analyse the direct and indirect impacts of environmental contexts and individual characteristics on rural residents' EAs. The results showed the environmental context can indirectly impact EAs by amplifying the influence of individual characteristics such as family structure and income on EAs. The EAs are influenced by different local environmental contexts among the east, middle and west of China's drylands. The humidity attitude was influenced by precipitation only in the highly arid western and middle regions, while precipitation attitude is strongly influenced by land surface temperature and humidity in eastern China's drylands. These findings hold important implications for understanding the cross-scale impact of environmental contexts on EAs in drylands.

In situ/Operando Synchrotron Radiation Analytical Techniques for CO2/CO Reduction Reaction: From Atomic Scales to Mesoscales.

Electrocatalytic carbon dioxide/carbon monoxide reduction reaction (CO(2)RR) has emerged as a prospective and appealing strategy to realize carbon neutrality for manufacturing sustainable chemical products. Developing highly active electrocatalysts and stable devices has been demonstrated as effective approach to enhance the conversion efficiency of CO(2)RR. In order to rationally design electrocatalysts and devices, a comprehensive understanding of the intrinsic structure evolution within catalysts and micro-environment change around electrode interface, particularly under operation conditions, is indispensable. Synchrotron radiation has been recognized as a versatile characterization platform, garnering widespread attention owing to its high brightness, elevated flux, excellent directivity, strong polarization and exceptional stability. This review systematically introduces the applications of synchrotron radiation technologies classified by radiation sources with varying wavelengths in CO(2)RR. By virtue of in situ/operando synchrotron radiationanalytical techniques, we also summarize relevant dynamic evolution processes from electronic structure, atomic configuration, molecular adsorption, crystal lattice and devices, spanning scales from the angstrom to the micrometer. The merits and limitations of diverse synchrotron characterization techniques are summarized, and their applicable scenarios in CO(2)RR are further presented. On the basis of the state-of-the-art fourth-generation synchrotron facilities, a perspective for further deeper understanding of the CO(2)RR process using synchrotron radiation analytical techniques is proposed.

Nanomechanical behavior of coal with heterogeneous minerals and pores using nanoindentation.

The coal's mechanical properties have a significant influence on mining safety and the mine environment. Preparing a standard sample and conducting repeat mechanical testing are challenging because the coal is primarily soft, fragmented, and rich in developed fractures. This study used nanoindentation technology, combined with X-ray diffraction, small-angle X-ray, a high magnification microscope, and mechanical parameter scale-up analysis, to study the micromechanical of three coals being dominated by heterogeneous components and pores. The results show that load-displacement curves with different maximum loads (50 mN, 100 mN, and 200 mN) all appear the pop-in events, and coal heterogeneity affects the frequency of their occurrence. As the maximum load is increased, pop-in event of DSC appears once each, YW increases from zero to three times and HM decreases from four to two times. The heterogeneity of pore structure has little effect on residual displacement, which is mainly affected by hard minerals, and hard minerals reduce the law that residual displacement increases with the increase in maximum load. The micromechanical parameters of soft coals are mainly affected by large pores, while hard coals are mainly affected by hard minerals. The coal's heterogeneity does not affect the linear relationship between hardness and elastic modulus, but stronger heterogeneity will weaken the linear relationship between fracture toughness and elastic modulus. Compared to the mechanical parameters after scale-up, the values obtained based on nanoindentation are less than 15.588% larger, and the increase in the heterogeneity and hard minerals can make the predicted parameters more accurate. The nanoindentation technique can not only provide an efficient and accurate method for studying the mechanical properties of heterogeneous coal at the nanoscale, an important guide for large-scale coal.

Overview of Multi-Scale Simulation Techniques for Three Typical Steel Manufacturing Processes.

Steel products typically undergo intricate manufacturing processes, commencing from the liquid phase, with casting, hot rolling, and laminar cooling being among the most crucial processes. In the background of carbon neutrality, thin-slab casting and direct rolling (TSCR) technology has attracted significant attention, which integrates the above three processes into a simpler and more energy-efficient sequence compared to conventional methods. Multi-scale computational modeling and simulation play a crucial role in steel design and optimization, enabling the prediction of properties and microstructure in final steel products. This approach significantly reduces the time and cost of production compared to traditional trial-and-error methodologies. This study provides a review of cross-scale simulations focusing on the casting, hot-rolling, and laminar cooling processes, aiming at presenting the key techniques for realizing cross-scale simulation of the TSCR process.

Study on the Cross-Scale Effects of Microscopic Interactions and Mechanical Properties of Rigid Polyurethane Foam Driven by Negative-Temperature Environments.

In order to investigate the cross-scale effects of the interaction between the hard and soft segments of stiff polyurethane foam on the material's mesoscopic pore structure and macroscopic compression characteristics in various negative-temperature environments, this paper used molecular dynamics to calculate the interaction differences between hard and soft segments in different negative-temperature environments. The effects of various negative-temperature settings on the cell structure of stiff polyurethane foam were investigated using scanning electron microscopy and Image J software. Finally, macro experiments were used to determine the influence of a negative-temperature environment on the characteristics of stiff polyurethane foam (such as compressibility). The molecular simulation calculation results show that in a negative-temperature environment, decreasing temperature gradually increases the interaction between hard segment molecules and soft segment molecules, resulting in an increase in the molecules' modulus and cohesive energy density. The scanning electron microscope results reveal that a negative-temperature environment gradually increases the pore diameter of stiff polyurethane foam. The compression experiment findings demonstrate that, for the same service duration, the compressive strength in the -20 °C environment is 27.53% higher than that in the 0 °C environment. The study's findings reveal a microscopic mechanism for the following receiving alterations and toughness enhancement of rigid polyurethane foam throughout service in negative-temperature conditions.

ESF-YOLO: an accurate and universal object detector based on neural networks.

As an excellent single-stage object detector based on neural networks, YOLOv5 has found extensive applications in the industrial domain; however, it still exhibits certain design limitations. To address these issues, this paper proposes Efficient Scale Fusion YOLO (ESF-YOLO). Firstly, the Multi-Sampling Conv Module (MSCM) is designed, which enhances the backbone network's learning capability for low-level features through multi-scale receptive fields and cross-scale feature fusion. Secondly, to tackle occlusion issues, a new Block-wise Channel Attention Module (BCAM) is designed, assigning greater weights to channels corresponding to critical information. Next, a lightweight Decoupled Head (LD-Head) is devised. Additionally, the loss function is redesigned to address asynchrony between labels and confidences, alleviating the imbalance between positive and negative samples during the neural network training. Finally, an adaptive scale factor for Intersection over Union (IoU) calculation is innovatively proposed, adjusting bounding box sizes adaptively to accommodate targets of different sizes in the dataset. Experimental results on the SODA10M and CBIA8K datasets demonstrate that ESF-YOLO increases Average Precision at 0.50 IoU (AP50) by 3.93 and 2.24%, Average Precision at 0.75 IoU (AP75) by 4.77 and 4.85%, and mean Average Precision (mAP) by 4 and 5.39%, respectively, validating the model's broad applicability.

Study on the Melt Rheological Characterization of Micro-Tube Gas-Assisted Extrusion Based on the Cross-Scale Viscoelastic Model.

In the micro-tube gas-assisted extrusion process, flow theories ignoring cross-scale viscoelastic variations fail to effectively characterize the rheological state of the melt. To investigate the impact of cross-scale viscoelastic variation on the quality of the micro-tube gas-assisted extrusion, a 3D multiphase flow extrusion model incorporating a double gas-assisted layer was developed. Subsequently, we modified the DCPP constitutive equations based on the cross-scale factor model. Both the traditional and gas-assisted extrusions were simulated under macroscale and cross-scale models using the Ansys Polyflow. Finally, using the established gas-assisted extrusion platform, extrusion experiments were conducted. The results indicate that, owing to the reduced melt viscosity under the cross-scale model, the deformation behavior of the melt is more pronounced than in the macroscale model. The cross-scale model's numerical results more closely match the experimental outcomes under the same parameters, thereby confirming the feasibility of the theoretical analysis and numerical simulation. Moreover, the predictive capability of the cross-scale model for the micro-tube gas-assisted extrusion is further validated through numerical and experimental analyses with varying parameters. It is demonstrated that the cross-scale viscoelastic variation is a critical factor that cannot be overlooked in the gas-assisted extrusion.

Strong, Tough, and Biocompatible Poly(vinyl alcohol)-Poly(vinylpyrrolidone) Multiscale Network Hydrogels Reinforced by Aramid Nanofibers.

Poly(vinyl alcohol) (PVA) hydrogels are water-rich, three-dimensional (3D) network materials that are similar to the tissue structure of living organisms. This feature gives hydrogels a wide range of potential applications, including drug delivery systems, articular cartilage regeneration, and tissue engineering. Due to the large amount of water contained in hydrogels, achieving hydrogels with comprehensive properties remains a major challenge, especially for isotropic hydrogels. This study innovatively prepares a multiscale-reinforced PVA hydrogel from molecular-level coupling to nanoscale enhancement by chemically cross-linking poly(vinylpyrrolidone) (PVP) and in situ assembled aromatic polyamide nanofibers (ANFs). The optimized ANFs-PVA-PVP (APP) hydrogels have a tensile strength of ≈9.7 MPa, an elongation at break of ≈585%, a toughness of ≈31.84 MJ/m3, a compressive strength of ≈10.6 MPa, and a high-water content of ≈80%. It is excellent among all reported PVA hydrogels and even comparable to some anisotropic hydrogels. System characterizations show that those performances are attributed to the particular multiscale load-bearing structure and multiple interactions between ANFs and PVA. Moreover, APP hydrogels exhibit excellent biocompatibility and a low friction coefficient (≈0.4). These valuable performances pave the way for broad potential in many advanced applications such as biological tissue replacement, flexible wearable devices, electronic skin, and in vivo sensors.

Cross-Modality Reference and Feature Mutual-Projection for 3D Brain MRI Image Super-Resolution.

High-resolution (HR) magnetic resonance imaging (MRI) can reveal rich anatomical structures for clinical diagnoses. However, due to hardware and signal-to-noise ratio limitations, MRI images are often collected with low resolution (LR) which is not conducive to diagnosing and analyzing clinical diseases. Recently, deep learning super-resolution (SR) methods have demonstrated great potential in enhancing the resolution of MRI images; however, most of them did not take the cross-modality and internal priors of MR seriously, which hinders the SR performance. In this paper, we propose a cross-modality reference and feature mutual-projection (CRFM) method to enhance the spatial resolution of brain MRI images. Specifically, we feed the gradients of HR MRI images from referenced imaging modality into the SR network to transform true clear textures to LR feature maps. Meanwhile, we design a plug-in feature mutual-projection (FMP) method to capture the cross-scale dependency and cross-modality similarity details of MRI images. Finally, we fuse all feature maps with parallel attentions to produce and refine the HR features adaptively. Extensive experiments on MRI images in the image domain and k-space show that our CRFM method outperforms existing state-of-the-art MRI SR methods.

Moving beyond the panarchy heuristic.

Panarchy is a heuristic of complex system change rooted in resilience science. The concept has been rapidly assimilated across scientific disciplines due to its potential to envision and address sustainability challenges, such as climate change and regime shifts, that pose significant challenges for humans in the Anthropocene. However, panarchy has been studied almost exclusively via qualitative research. Quantitative approaches are scarce and preliminary but have revealed novel insights that allow for a more nuanced understanding of the heuristic and resilience science more generally. In this roadmap we discuss the potential for future quantitative approaches to panarchy. Transdisciplinary development of quantitative approaches, combined with advances in data accrual, curation and machine learning, may build on current tools. Combined with qualitative research and traditional approaches used in ecology, quantification of panarchy may allow for broad inference of change in complex systems of people and nature and provide critical information for management of social-ecological systems.