Residual Map Research Articles

Efficient Light Field Transmission via Enhanced Resampling Reconstruction and Angular Attention Estimation

Light field (LF) imaging technology has significantly enhanced immersive virtual reality experiences and advanced computer vision tasks like depth estimation and 3D reconstruction. However, the massive data volume of light field images (LFIs) presents a transmission challenge. To address this issue, we propose a novel intelligent LF transmission approach integrating enhanced resampling reconstruction and angular attention estimation. The proposed enhanced resampling reconstruction method reduces spatial redundancy by downsampling before transmission and upsampling afterward, aided by a pre-calculated residual map to minimizing quality loss. To reduce angular redundancy while preserving perceptual quality, we designed a LF angular attention estimation network employing specifically designed angular attention kernels to guide differential transmission within the angular domain. To train this network, we construct the first LF eye-tracking dataset, featuring numerous of samples, diverse realistic scenes, and a wealth of data elements, particularly real user angular attention data, establishing it a key benchmark in LF transmission. Extensive experiments demonstrate that our transmission strategy decreases transmission time by 97.3% with only a 3.1% perceptual quality loss. Subjective experiments further confirm the superior performance of our method. An implementation of this paper and the constructed dataset are available at https://github.com/VincentQQu/LF-Transmission.

Utilizing of Geophysical Methods for Geothermal Exploration at Guru Kinayan Village, Tanah Karo Regency

Guru Kinayan, located at the foothills of mount Sinabung in Karo regency, is characterized by several geothermal manifestations, highlighting the necessity of studying its geothermal potential. This research aims to assess the geothermal prospects of the area through 2-D electrical resistivity, geomagnetic surveys, and (XRD) methods. To acquire 2-D resistivity data, a Wenner Schlumberger array with a 5-meter electrode spacing was utilized, and the data were processed using Res2Dinv. Geomagnetic surveys were conducted using a PPM, with the results analyzed using Mag2dc and Surfer 13 to determine susceptibility results. The analysis of rock composition was performed using XRD. The 2-D resistivity results indicated resistivity values ranging from 1 to 1250 Ωm. Values between 1 and 20 Ωm indicated as alluvium, indicating a reservoir for geothermal, while resistivity values >1000 indicates as limestone, which indicated as a heat conductor. The magnetic residual map shows geomagnetic values (20 to 380 nT), suggesting a geothermal source, and the susceptibility values ranging (0.0013 x 10³ to 0.0088 x 10³) indicates the presence of igneous rocks, specifically andesite lava and pyroclastic rocks. The XRD results show arsenopirit and quartz in the geothermal area. Based on the integrated results, the study area has significant geothermal potential.

The influence of feed molar flow rate on key streams properties and duties in ethanol-water azeotropic distillation process

Investigation on the influence of near azeotropic feed molar flow rate on properties of key streams connecting main process units of an ethanol-water azeotropic distillation pilot processis presented. The heterogeneous azeotropic distillation model with cyclohexane entrainer was configured in Aspen Plus® V10 commercial simulation software. Property analysis and prediction of plant’s performancewere achieved using Non-Random Two Liquid Redlich-Kwong thermodynamic model. Both residual curve maps and ternary diagrams were used to determineseparation possibility and presence of ethanol-water azeotrope in the formed ternary mixtureduring distillation synthesis. The process convergence was addressed using flowsheet convergence and balance node. Distillation of ethanol- water azeotropic mixture was simulated at constant recycle ratio (R = 5), numberof stages (N = 12) and pressure (P = 1 atm). Data obtained were analysed in MS Excel 2013. Results showed that simultaneous increase ofmolar flow rate of the near azeotropic feed and feed ethanol concentration improves ethanol purity but results in retrograde phenomenon which may humper plant’s performance and increase stream flow rates and therefore increase energy duties. The condenser duty of the main column wasaround 90.94% higher than the recycle column’scondenser duty. Also, reboiler duty of the main column were nearly 90.25% higher than the recycle column’s reboiler duty.It was concluded that setting 18 to 20 kmol/h feed molar flow rate enhances higher energy efficiency and improves ethanol product purity. Careful distillation synthesis and plant’s monitoring are recommended to improve ethanol-water azeotropic distillation plant’s performance and address retrograde phenomena.

Nucleotide-resolution Mapping of RNA N6-Methyladenosine (m6A) modifications and comprehensive analysis of global polyadenylation events in mRNA 3' end processing in malaria pathogen Plasmodium falciparum.

Plasmodium falciparum is an obligate human parasite of the phylum Apicomplexa and is the causative agent of the most lethal form of human malaria. Although N6-methyladenosine modification is thought to be one of the major post-transcriptional regulatory mechanisms for stage-specific gene expression in apicomplexan parasites, the precise base position of m6A in mRNAs or noncoding RNAs in these parasites remains unknown. Here, we report global nucleotide-resolution mapping of m6A residues in P. falciparum using DART-seq technology, which quantitatively displayed a stage-specific, dynamic distribution pattern with enrichment near mRNA 3' ends. In this process we identified 894, 788, and 1,762 m6A-modified genes in Ring, Trophozoite and Schizont stages respectively, with an average of 5-7 m6A sites per-transcript at the individual gene level. Notably, several genes involved in malaria pathophysiology, such as KAHRP, ETRAMPs, SERA and stress response genes, such as members of Heat Shock Protein (HSP) family are highly enriched in m6A and therefore could be regulated by this RNA modification. Since we observed preferential methylation at the 3' ends of P. falciparum transcripts and because malaria polyadenylation specificity factor PfCPSF30 harbors an m6A reader 'YTH' domain, we reasoned that m6A might play an important role in 3'-end processing of malaria mRNAs. To investigate this, we used two complementary high-throughput RNA 3'-end mapping approaches, which provided an initial framework to explore potential roles of m6A in the regulation of alternative polyadenylation (APA) during malaria development in human hosts.

Symmetric deformable registration of multimodal brain magnetic resonance images via appearance residuals.

Deformable registration of multimodal brain magnetic resonance images presents significant challenges, primarily due to substantial structural variations between subjects and pronounced differences in appearance across imaging modalities. Here, we propose to symmetrically register images from two modalities based on appearance residuals from one modality to another. Computed with simple subtraction between modalities, the appearance residuals enhance structural details and form a common representation for simplifying multimodal deformable registration. The proposed framework consists of three serially connected modules: (i) an appearance residual module, which learns intensity residual maps between modalities with a cycle-consistent loss; (ii) a deformable registration module, which predicts deformations across subjects based on appearance residuals; and (iii) a deblurring module, which enhances the warped images to match the sharpness of the original images. The proposed method, evaluated on two public datasets (HCP and LEMON), achieves the highest registration accuracy with topology preservation when compared with state-of-the-art methods. Our residual space-guided registration framework, combined with GAN-based image enhancement, provides an effective solution to the challenges of multimodal deformable registration. By mitigating intensity distribution discrepancies and improving image quality, this approach improves registration accuracy and strengthens its potential for clinical application.

DeepMoIC: multi-omics data integration via deep graph convolutional networks for cancer subtype classification

BackgroundAchieving precise cancer subtype classification is imperative for effective prognosis and treatment. Multi-omics studies, encompassing diverse data modalities, have emerged as powerful tools for unraveling the complexities of cancer. However, owing to the intricacies of biological data, multi-omics datasets generally show variations in data types, scales, and distributions. These intractable problems lead to challenges in exploring intact representations from heterogeneous data, which often result in inaccuracies in multi-omics information analysis.ResultsTo address the challenges of multi-omics research, our approach DeepMoIC presents a novel framework derived from deep Graph Convolutional Network (GCN). Leveraging autoencoder modules, DeepMoIC extracts compact representations from omics data and incorporates a patient similarity network through the similarity network fusion algorithm. To handle non-Euclidean data and explore high-order omics information effectively, we design a Deep GCN module with two strategies: residual connection and identity mapping. With extracted higher-order representations, our approach consistently outperforms state-of-the-art models on a pan-cancer dataset and 3 cancer subtype datasets.ConclusionThe introduction of Deep GCN shows encouraging performance in terms of supervised multi-omics feature learning, offering promising insights for precision medicine in cancer research. DeepMoIC can potentially be an important tool in the field of cancer subtype classification because of its capacity to handle complex multi-omics data and produce reliable classification findings.

Convergence of Damped Polarization Schemes for the FFT‐Based Computational Homogenization of Inelastic Media With Pores

ABSTRACTPorous microstructures represent a challenge for the convergence of FFT‐based computational homogenization methods. In this contribution, we show that the damped Eyre–Milton iteration is linearly convergent for a class of nonlinear composites with a regular set of pores, provided the damping factor is chosen between zero and unity. First, we show that an abstract fixed‐point method with non‐expansive fixed‐point operator and non‐trivial damping converges linearly, provided the associated residual mapping satisfies a monotonicity condition on a closed subspace. Then, we transfer this result to the framework of polarization schemes and conclude the linear convergence of the damped Eyre–Milton scheme for porous materials. We present general arguments which apply to a class of nonlinear composites and mixed stress‐strain loadings, as well. We show that the best contraction estimate is reached for a damping factor of , that is, for the polarization scheme of Michel–Moulinec–Suquet, and derive the corresponding optimal reference material. Our results generalize the recent work of Sab and co‐workers who showed that an adaptively damped Eyre–Milton scheme leads to linear convergence for a class of linear composites with pores. Finally, we report on computational experiments which support our findings.

The Spectrum Difference Enhanced Network for Hyperspectral Anomaly Detection

Most deep learning-based hyperspectral anomaly detection (HAD) methods focus on modeling or reconstructing the hyperspectral background to obtain residual maps from the original hyperspectral images. However, these methods typically do not pay enough attention to the spectral similarity in the complex environment, resulting in inadequate distinction between background and anomalies. Moreover, some anomalies and background are different objects, but they are sometimes recognized as the objects with the same spectrum. To address the issues mentioned above, this paper proposes a Spectrum Difference Enhanced Network (SDENet) for HAD, which employs variational mapping and Transformer to amplify spectrum differences. The proposed network is based on the encoder–decoder structure, which contains a CSWin-Transformer encoder, Variational Mapping Module (VMModule), and CSWin-Transformer decoder. First, the CSWin-Transformer encoder and decoder are designed to supplement image information by extracting deep and semantic features, where a cross-shaped window self-attention mechanism is designed to provide strong modeling capability with minimal computational cost. Second, in order to enhance the spectral difference characteristics between anomalies and background, a randomly sampling VMModule is presented for feature space transformation. Finally, all fully connected mapping operations are replaced with convolutional layers to reduce the model parameters and computational load. The effectiveness of the proposed SDENet is verified on three datasets, and experimental results show that it achieves better detection accuracy and lower model complexity compared with existing methods.

GACT-PPIS: Prediction of protein-protein interaction sites based on graph structure and transformer network

The prediction of protein-protein interaction sites (PPIS) is currently crucial for regulating many biological activities in cells and developing drugs for various diseases. Deep learning-based methods have been proposed for predicting PPIS, significantly reducing the manpower and time costs associated with traditional experimental methods such as yeast two-hybrid, mass spectrometry, and affinity purification. However, the predictive accuracy of these deep learning methods has not yet reached the expected level. Therefore, we introduce a model called GACT-PPIS.The design of the GACT-PPIS algorithm aims to utilize combined information from protein sequences and structures as input to predict protein-protein interaction sites. The core of GACT-PPIS utilizes an Enhanced Graph Attention Network (EGAT) with initial residual and identity mappings, along with a deep Transformer network as the basic units, supplemented by Graph Convolutional Networks (GCN), effectively aggregating information from neighboring nodes for each node. After multiple network layers, the information of the entire protein is also fused into the nodes, and the Transformer network further enhances the model's performance.Experimental results show that GACT-PPIS outperforms the most representative models in terms of Recall, F1-measure, MCC, AUROC, and AUPRC on the benchmark test set (Test-60). Additionally, on other independent test sets (UBTest-31-6), GACT-PPIS leads in terms of Accuracy, Precision, Recall, F1-measure, MCC, AUROC, and AUPRC compared to the most representative models. It is worth noting that GACT-PPIS demonstrates excellent generalization and versatility across different test sets, showcasing good performance on multiple test sets for the same trained GACT-PPIS model.

LGS-PPIS: A Local-Global Structural Information Aggregation Framework for Predicting Protein-Protein Interaction Sites.

Exploring protein-protein interaction sites (PPIS) is of significance to elucidating the intrinsic mechanisms of diverse biological processes. On this basis, recent studies have applied deep learning-based technologies to overcome the high cost of wet experiments for PPIS determination. However, the existing methods still suffer from two limitations that remain to be solved. Firstly, the process of feature aggregation in most methods only took into account node features, but ignored the complex edge features of the target residue to its neighbor residues, resulting in insufficient local feature extraction. Secondly, such feature aggregation was limited to aggregating spatially adjacent residues, and could not capture the "remote" residues that played a critical role in determining PPIS, which can be summed up as the lack of global feature at the residue level. To break the above limitations, a local-global structural information aggregation framework, LGS-PPIS, was proposed in this study, including two modules of edge-aware graph convolutional network (EA-GCN) and self-attention integrated with initial residual and identity mapping (SA-RIM), which achieved the aggregation of local and global information for PPIS prediction. Evaluation results of LGS-PPIS showed that the proposed method outperformed state-of-the-art deep learning methods on three widely used PPIS prediction benchmarks. Besides, the results of ablation experiments demonstrated that the local features from spatially adjacent residues and global features from "remote" residues separately captured by EA-GCN and SA-RIM could benefit the model performance. Among them, the former was shown to have a more significant role in the PPIS prediction.

Geophysical and geospatial insights into surface and subsurface characteristics for groundwater potential analysis, Ras Sudr, West Sinai, Egypt

Ras Sudr has garnered significant interest owing to various initiatives aimed at expanding and developing this area. It holds considerable strategic importance, serving as a key development hub in Sinai and a focal point for tourism destination. Moreover, it has a remarkable event with flashfloods which can be utilized for groundwater or direct usage. Integrating geophysical and geospatial analyses to study the surface and subsurface characteristics as well as identify groundwater potential areas in Ras Sudr, west-central Sinai is the main objective of this study. Firstly, geophysical data including gravity and magnetic methods have a crucial importance in qualitative and quantitative interpretation of the subsurface elements. Filtering techniques were implemented to distinguish between regional and residual anomalies. Geophysical data were subjected to the radial average power spectrum technique and 3D Euler deconvolution to identify the depth of the subsurface sources. The structure pattern that characterizes the interested region was defined by employing bandpass filter and edge detection appoarches using residual anomaly maps, regional anomaly maps, tilt derivatives and total horizontal gradient maps reflecting four distinct structural trends; NW-SE parallel to the Gulf of Suez, NEE-SWW parallel to the Syrian arc system, N-S parallel to the the Nile Valley, and NNE-SSW parallel to the Gulf of Aqaba. A basement relief map was constructed using 3D magnetic modelling showing that the depth of the basement ranges from 1.6 to 6.3 km. Secondly, the remote sensing data including Sentienl-2 and SRTM datasets were employed to extract the surface analyses in the GIS environment to develop the occurrence of groundwater potentiality utilizing seven factors; LU-LC, soil, geology, slope, drainage network and lineament density and rainfall data of the study area which were ranked from 1 to 5 and weighted according to their effective contribution to the infiltration of groundwater using AHP-GIS based multi criteria method. The resulting Groundwater Potential zone (GWPZ) was categorized into five zones from poor to excellent and validated using 41 observed wells. A significant quantity of wells was identified in the areas of high potential located to the west of the region, while five wells were situated within the moderate potential zone. Thereby, the GWPZ map identifying locations with viable groundwater resources suitable for habitation, development and economic plans in Sinai for decision makers.

Mind the Kinematics Simulation of Planet–Disk Interactions: Time Evolution and Numerical Resolution

Planet–disk interactions can produce kinematic signatures in protoplanetary disks. While recent observations have detected non-Keplerian gas motions in disks, their origins are still being debated. To explore this, we conduct 3D hydrodynamic simulations using the code FARGO3D to study nonaxisymmetric kinematic perturbations at two scale heights induced by Jovian planets in protoplanetary disks, followed by examinations of detectable signals in synthetic CO emission line observations at millimeter wavelengths. We advocate for using residual velocity or channel maps, generated by subtracting an azimuthally averaged background of the disk, to identify planet-induced kinematic perturbations. We investigate the effects of two basic simulation parameters, simulation duration and numerical resolution, on the simulation results. Our findings suggest that a short simulation (e.g., 100 orbits) is insufficient to establish a steady velocity pattern given our chosen viscosity (α = 10−3) and displays plenty of fluctuations on an orbital timescale. Such transient features could be detected in observations. By contrast, a long simulation (e.g., 1000 orbits) is required to reach steady state in kinematic structures. At 1000 orbits, the strongest detectable velocity structures are found in the spiral wakes close to the planet. Through numerical convergence tests, we find hydrodynamics results converge in spiral regions at a resolution of 14 cells per disk scale height or higher. Meanwhile, synthetic observations produced from hydrodynamic simulations at different resolutions are indistinguishable with 0.″1 angular resolution and 10 hr of integration time on Atacama Large Millimeter/submillimeter Array.

Outflow from the very massive Wolf-Rayet binary Melnick 34

Melnick 34 (Mk 34) is one of the most massive binary systems known and is one of the brightest X-ray point sources in the 30 Doradus region. We investigated the impact of this massive system on the surrounding interstellar medium (ISM) using the optical spectroscopic capabilities of the narrow-field mode (NFM) of the Multi-Unit Spectroscopic Explorer (MUSE). MUSE-NFM spatially resolved the ISM in the vicinity of Mk34 with a resolution comparable to that of the HST. The analysis of the [NII] λ 6583 and [SII] λ 6717 emission lines reveals a cone-like structure apparently originating from Mk 34 and extending southeast. Electron density maps and radial velocity measurements of the ISM lines further support an outflow scenario traced by these emissions. While no clear northwestern counterpart to this outflow was observed, we note increased extinction in that direction, towards the R136 cluster. The ISM material along the projected diagonal of the outflow on both sides of Mk 34 shows similar properties in terms of the emission line ratios seen in the Baldwin-Phillips-Terlevich diagram. These results are consistent across two observational epochs. Additionally, we examined the residual maps within a 0.5″ radius of Mk 34 after modeling and subtracting the point spread function. The observed variations in the residuals could potentially be linked to Mk 34’s known periodic behavior. However, further observations with appropriate cadence are needed to fully monitor the 155 day periodicity of Mk 34’s X-ray emissions.

Multiple Upward Residual Gravity and Magnetic Maps to Estimate the Deep Crustal Layers and Moho Discontinuity in Eastern Iraq

The upward continuation technique was applied to the gravity and magnetic data to obtain the residual anomalies maps in eastern Iraq. A new processing method is introduced to calculate the crustal layers depending on the relation between the total range of the residual anomalies of gravity, magnetic data and the upward continuation, with multi-elevations. The significant relationship shows slope variations, which may reflect the maximum contrast in the physical properties (density and/or susceptibility) and the intersection point representing the depth of layer contacts. This method was applied to the eastern central part of Iraq, adjacent to the Iranian borders. It is found that the average depth range to the Jurassic formations, top basement rocks, transition layer (Curie point), and Moho discontinuity are 3.1- 4.8 km, 8-12.9 km, 20.9-22.4 km and 43.6 km, respectively. The Sedimentary cover in the study area ranges from 8 to 13 km. The maximum depth to the Moho discontinuity obtained by gravity data in the study area is 43.6 km. The radially average power spectrum technique was applied in the study area to calculate the gravity depth and magnetic sources' depth. Three depth levels of sources from gravity and magnetic data were obtained. These depths are shallow sources, average basement rock and deep sources (Moho) at 3km, 13.2-12.1 km, and 32.2-34km, respectively. The obtained results concerning the main layers in the crust were compared between the two methods, and the new method shows more reliable results. The detected depth levels refer to top Jurassic, basement rocks, and Moho discontinuity. The new method is believed to be used to detect the crustal layers depth with good results.

Estimation of Subsurface Structure Using Euler Deconvolution Method of Magnetic Data at the Geothermal Area of Sonai Village and its Surroundings, Konawe Regency

It has been carried out research with the aim of determining the subsurface structure at the geothermal area of Sonai Village and its surroundings, Konawe Regency. The data used are magnetic data obtained through field measurements at 126 points in the N180oS direction. After the data were subjected to diurnal and IGRF corrections, a residual (local) magnetic field anomaly of around -150 nT to 90 nT was obtained. On the residual magnetic anomaly map which has been reduced to the Pole (RTP), the Euler Deconvolution (ED) method is applied to the Index Structure N=0 to estimate the subsurface structure in the form of the presence of minor faults, and it is known that there are 5 minor faults at a depth of around 9 to 38 meters. Information on the existence of these faults is then used in 2D modeling. Modeling results show that these minor faults cut through two rock layers, which are the layers composed of conglomerate rocks from the Alangga Formation and peridotites as bedrock from the Ultramafic Complex. One of the minor faults closest to the manifestation area (hot spring) is at coordinates around 4o1’12.412” S and 122o7’24.263” E to 4o1’15.532” S and 122o7’19.561” E with a distance of ±15 meters. The existence of these minor faults is thought to be the migration routes for heat flow or conduction to the surface at the geothermal area of Sonai Village and its surroundings.

Unveiling image source: Instance-level camera device linking via context-aware deep Siamese network

Unveiling the source of an image is one of the most effective ways to validate the originality, authenticity, and reliability in the field of digital forensics. Source camera device identification can identify the specific camera device used to take a photo under investigation. While great progress has been made by the photo-response non-uniformity (PRNU)-based methods over the past decade, the challenge of instance-level source camera device linking, which verifies whether two images in question were captured by the same camera device, remains significant. This challenge is mainly due to the absence of auxiliary images to construct a clean camera fingerprint for each camera, particularly dealing with small image sizes. To overcome this limitation, in this paper, we formulate the task of source device linking as a binary classification problem and propose a simple yet effective framework based on a context-aware deep Siamese network. We take advantage of a Siamese architecture to extract the intrinsic camera device-related noise patterns from a pair of image patches in parallel for comparisons without any auxiliary images. Moreover, a recurrent criss-cross group is utilized to aggregate contextual information in the noise residual maps to alleviate the problem that PRNU noise maps are easily contaminated by the additive noises from image contents. For reliable device linking, we employ a patch-selection strategy on a pair of test images to adaptively choose suitable image patch pairs according to image contents. The final decision of a pair of test images is obtained from the average similarity score of the selected image patch pairs. Compared with existing state-of-the-art methods, our proposed framework can achieve better performance on both the tasks of source camera identification and source device linking without any prior knowledge, i.e., reliable camera fingerprints, regardless of whether the camera devices are “seen” or “unseen” in the training stage. The experimental results on two standard image forensic datasets demonstrate that the proposed method not only shows robustness with respect to different image patch sizes and image quality degenerations, but also has a generalization ability across digital camera and smartphone devices.

Inhomogeneous strain partitioning in the Cenozoic Bohai Bay Basin controlled by pre-existing crustal fabrics and oblique subduction: Insights from the Jizhong and Huanghua subbasins

The tectonic features and geodynamics of the Bohai Bay Basin (BBB) have been extensively studied over several decades. It is generally accepted that the Cenozoic BBB was subjected to a superimposed extensional and strike-slip stress field. However, the specific basin-forming mechanism remains ambiguous. Basin-scale studies with high spatial and temporal resolution reveal how strain migrates and localizes during multiphase rifting. Understanding these strain-related processes is crucial for deciphering the formation mechanisms and controlling factors of rift basins. This study investigates the rift-related evolution and strain partitioning in the Jizhong and Huanghua subbasins, the western half of the BBB. Using high-resolution 3D and 2D seismic datasets and well data, we present detailed geological sections, residual thickness maps, and fault systems, showing two distinct patterns of strain partitioning in the Cenozoic. In the Jizhong Subbasin, strain migrated inwards from the NE-trending boundary faults and finally became localized along the NE-trending rift axis; in the Huanghua Subbasin, strain migrated northeastwards from the southern part and then was largely localized along the near E–W-trending faults in the northeastern part. By integrating our results with previous studies of other subbasins, we identify two separate extensional domains within the Cenozoic BBB. The suborthogonal extensional domain in the Jizhong Subbasin is dominated by relatively stable NW–SE extension. In contrast, the extended areas to the east (i.e., the Huanghua, Bozhong, and Liaohe-Liaodongwan subbasins) comprise the oblique extensional domain, characterized by asymmetric transtensional pull-apart deformation and subjected to a clockwise-rotated extensional stress field from NW–SE to NNW–SSE. We suggest that the spatially inhomogeneous strain partitioning in the BBB has developed since the middle Eocene, and was controlled by the interaction of pre-existing crustal fabrics (e.g., the Tan-Lu Fault Zone) and the oblique subduction of the Pacific Plate. The boundary between the two domains may represent a transition zone where the margin-parallel residual velocity component of the oblique convergence decreased considerably landwards. Our study highlights the Cenozoic strain evolution and the associated geodynamics in the BBB, emphasizing the strain complexity within the back-arc rift basin under the oblique subduction background.

Comprehensive Graph Gradual Pruning for Sparse Training in Graph Neural Networks.

Graph neural networks (GNNs) tend to suffer from high computation costs due to the exponentially increasing scale of graph data and a large number of model parameters, which restricts their utility in practical applications. To this end, some recent works focus on sparsifying GNNs (including graph structures and model parameters) with the lottery ticket hypothesis (LTH) to reduce inference costs while maintaining performance levels. However, the LTH-based methods suffer from two major drawbacks: 1) they require exhaustive and iterative training of dense models, resulting in an extremely large training computation cost, and 2) they only trim graph structures and model parameters but ignore the node feature dimension, where vast redundancy exists. To overcome the above limitations, we propose a comprehensive graph gradual pruning framework termed CGP. This is achieved by designing a during-training graph pruning paradigm to dynamically prune GNNs within one training process. Unlike LTH-based methods, the proposed CGP approach requires no retraining, which significantly reduces the computation costs. Furthermore, we design a cosparsifying strategy to comprehensively trim all the three core elements of GNNs: graph structures, node features, and model parameters. Next, to refine the pruning operation, we introduce a regrowth process into our CGP framework, to reestablish the pruned but important connections. The proposed CGP is evaluated over a node classification task across six GNN architectures, including shallow models [graph convolutional network (GCN) and graph attention network (GAT)], shallow-but-deep-propagation models [simple graph convolution (SGC) and approximate personalized propagation of neural predictions (APPNP)], and deep models [GCN via initial residual and identity mapping (GCNII) and residual GCN (ResGCN)], on a total of 14 real-world graph datasets, including large-scale graph datasets from the challenging Open Graph Benchmark (OGB). Experiments reveal that the proposed strategy greatly improves both training and inference efficiency while matching or even exceeding the accuracy of the existing methods.

Volcanic weathered crust reservoir analog: Insights from fault-controlled fracture permeability in the trachytic rocks of the Pernambuco Basin, NE Brazil

Naturally fractured volcanic rocks represent a globally significant potential hydrocarbon reservoirs and can also be used for CO2 storage; however, the study of volcanic reservoirs in Brazil is limited. Although numerous volcanic reservoir types have been classified, relatively few studies are dedicated to the weathering of crust-like reservoirs, i.e., volcanic rocks undergoing subaerially alteration by weathering prior to deep burial. This study aims to characterize a volcanic weathered crust reservoir analog located in the coastal zone of the Pernambuco Basin, NE Brazil. We applied a multiscale approach motivated by the expressive potential of volcanic rocks as reservoirs in this marginal basin and considered their implications for CO2 geological storage projects. The deformational aspects observed in this analog have also been observed in other volcanic rocks that crop out in the coastal zone, thus, the data provided here can guide future characterization of both onshore and offshore reservoirs. We focused on an 80 m wide outcrop and employed geometric and topological analysis of two-dimensional fracture networks to investigate the influence of regional fault zone control on the fracture permeability tensors. The parameters of fracture systems were studied using terrestrial gravimetric data, orthophoto mosaic obtained by an unmanned aerial vehicle (UAV), petrography, and outcrop-based structural data. The investigated outcrop includes porphyritic trachytes with euhedral sanidine crystals and minor quartz phenocrysts, sourced from the Ipojuca Magmatic Suite. The residual Bouguer map indicates the influence of at least two regional rift fault zones, i.e., NW-SE (strike-slip) and NE-SW (normal) trending systems, in the studied rocks. These trachytic rocks exhibited a marked dissolution of sanidine phenocrysts and matrix, columnar disjunctions, and natural fractures, causing the porosity and permeability of the reservoir analog to increase. Geometric and topological analyses showed that the south-central sector of the outcrop exhibits heightened connectivity within the fracture network. Furthermore, the principal fracture permeability is controlled by the primary families of the fracture network and regional fault zones. Our results demonstrate that the fracture network observed in the volcanic reservoir analog was influenced by the post-rift tectonic reactivation of Cretaceous structures within the Pernambuco Basin. The secondary porosity resulting from weathering and subsequent dissolution of the primary constituents and fractures of the rocks studied demonstrates the potential for the exploration of these reservoirs as models for known occurrences of volcanic rocks in the onshore and offshore domains of the Pernambuco Basin.

Novel direct injection electro-hydraulic model-based controller for high efficiency internal combustion engines

During the past years, automotive industries developed several technologies suitable to increase efficiency and reduce emissions from Internal Combustion Engines (ICEs). Among them, the adoption of high-pressure injection systems is considered crucial to optimize air-fuel mixture formation. However, the use of these technologies also promotes the formation of particulate matter (PM Particulate Matter), which is a direct result of charge stratification and fluid film on the cylinder walls. Therefore, to obtain a proper mixture formation without the risk of wall impingement, the utilization of consecutive injections is mandatory. Since modern Gasoline Direct Injection (GDI) systems are typically characterized by electrical-actuated injectors connected to a single high-pressure rail, a deep understanding of electrical and hydraulic effects among two close injection events becomes essential. This paper analyzes the combinations of electrical and hydraulic effects that occur in a high-pressure GDI system performing multiple injections. By using a specifically developed open vessel flushing bench, the injection system has been characterized in terms of pressure wave propagation as well as electrical distortions of the driving current profile of the injectors. The analysis of the experimental data has allowed for the calibration of the residual magnetization characteristic map in addition to the development of a pressure wave propagation control-oriented model. Finally, a Magnetization and Pressure Wave (MPW) correction strategy, easily implementable on an Electronic Control Unit (ECU) without the need for additional sensors, has been proposed. By running the MPW strategy, the error between the actual and expected injected mass has been reduced below 5% in all tested conditions.