Coarse-grained Method Research Articles

Shape-Aware Adversarial Learning for Scribble-Supervised Medical Image Segmentation with a MaskMix Siamese Network: A Case Study of Cardiac MRI Segmentation

The transition in medical image segmentation from fine-grained to coarse-grained annotation methods, notably scribble annotation, offers a practical and efficient preparation for deep learning applications. However, these methods often compromise segmentation precision and result in irregular contours. This study targets the enhancement of scribble-supervised segmentation to match the accuracy of fine-grained annotation. Capitalizing on the consistency of target shapes across unpaired datasets, this study introduces a shape-aware scribble-supervised learning framework (MaskMixAdv) addressing two critical tasks: (1) Pseudo label generation, where a mixup-based masking strategy enables image-level and feature-level data augmentation to enrich coarse-grained scribbles annotations. A dual-branch siamese network is proposed to generate fine-grained pseudo labels. (2) Pseudo label optimization, where a CNN-based discriminator is proposed to refine pseudo label contours by distinguishing them from external unpaired masks during model fine-tuning. MaskMixAdv works under constrained annotation conditions as a label-efficient learning approach for medical image segmentation. A case study on public cardiac MRI datasets demonstrated that the proposed MaskMixAdv outperformed the state-of-the-art methods and narrowed the performance gap between scribble-supervised and mask-supervised segmentation. This innovation cuts annotation time by at least 95%, with only a minor impact on Dice performance, specifically a 2.6% reduction. The experimental outcomes indicate that employing efficient and cost-effective scribble annotation can achieve high segmentation accuracy, significantly reducing the typical requirement for fine-grained annotations.

Gearbox Fault Diagnosis Based on Adaptive Variational Mode Decomposition–Stationary Wavelet Transform and Ensemble Refined Composite Multiscale Fluctuation Dispersion Entropy

Existing gearbox fault diagnosis methods are prone to noise interference and cannot extract comprehensive fault signals, leading to misdiagnosis or missed diagnosis. This paper proposes a method for gearbox fault diagnosis based on adaptive variational mode decomposition–stationary wavelet transform (AVMD-SWT) and ensemble refined composite multiscale fluctuation dispersion entropy (ERCMFDE). Initially, the kurtosis coefficient and autocorrelation coefficient are presented, and the Intrinsic Mode Functions are denoised through the application of AVMD-SWT. Secondly, the coarse-grained processing method of composite multiscale fluctuation dispersion entropy is extended to encompass three additional approaches: first-order central moment, second-order central moment, and third-order central moment. This enables the comprehensive extraction of feature information from the time series, thereby facilitating the formation of an initial hybrid feature set. Subsequently, recursive feature elimination (RFE) is employed for feature selection. Ultimately, the outcomes of the faults diagnoses are derived through the utilization of a Support Vector Machine with a Sparrow Search Algorithm (SSA-SVM), with the actual faults data collection and analysis conducted on an experimental platform for gearbox fault diagnosis. The experiments demonstrate that the method can accurately identify gearbox faults and achieve a high diagnostic accuracy of 98.78%.

A hybrid scaling coarse-graining method based on a computational fluid dynamics-discrete element method

A hybrid scaling coarse-graining method based on a computational fluid dynamics-discrete element method

Dissipation at limited resolutions: power law and detection of hidden dissipative scales

Abstract Non-equilibrium systems, in particular, living organisms, are maintained by irreversible transformations of energy that drive diverse functions. Quantifying their irreversibility, as measured by energy dissipation, is essential for understanding the underlying mechanisms. However, existing techniques usually overlook experimental limitations, either by assuming full information or by employing a coarse-graining method that requires knowledge of the structure behind hidden degrees of freedom. Here, we study the inference of dissipation from finite-resolution measurements by employing a recently developed model-free estimator that considers both the sequence of coarse-grained transitions and the waiting time distributions: σ 2 = σ 2 ℓ + σ 2 t . The dominant term σ 2 ℓ originates from the sequence of observed transitions. We find that it scales with resolution following a power law. Comparing the scaling exponent with a previous estimator highlights the importance of accounting for flux correlations at lower resolutions. σ 2 t comes from asymmetries in waiting time distributions. It is non-monotonic in resolution, with its peak position revealing characteristic scales of the underlying dissipative process, consistent with observations in the actomyosin cortex of starfish oocytes. Alternatively, the characteristic scale can be detected in a crossover of the scaling of σ 2 ℓ . This provides a novel perspective for extracting otherwise hidden characteristic dissipative scales directly from dissipation measurements. We illustrate these results in biochemical models as well as complex networks. Overall, this study highlights the significance of resolution considerations in non-equilibrium systems, providing insight into the interplay between experimental resolution, entropy production and underlying complexity.

Chemically Transferable Electronic Coarse Graining for Polythiophenes.

Recent advances in machine-learning-based electronic coarse graining (ECG) methods have demonstrated the potential to enable electronic predictions in soft materials at mesoscopic length scales. However, previous ECG models have yet to confront the issue of chemical transferability. In this study, we develop chemically transferable ECG models for polythiophenes using graph neural networks. Our models are trained on a data set that samples over the conformational space of random polythiophene sequences generated with 15 different monomer chemistries and three different degrees of polymerization. We systematically explore the impact of coarse-grained representation on ECG accuracy, highlighting the significance of preserving the C-β coordinates in thiophene. We also find that integrating unique polymer sequences into training enhances the model performance more efficiently than augmenting conformational sampling for sequences already in the training data set. Moreover, our ECG models, developed initially for one property and one level of quantum chemical theory, can be efficiently transferred to related properties and higher levels of theory with minimal additional data. The chemically transferable ECG model introduced in this work will serve as a foundation model for new classes of chemically transferable ECG predictions across chemical space.

Coarse-grained molecular dynamic model and wettability simulation of graphite materials

Although progress has been made in high-performance computing, there are still limitations on temporal and spatial scales of molecular dynamic calculations. A major issue in molecular dynamic (MD) simulations is the computational cost, and coarse-grained methods can save computational costs and accelerate calculations by reducing the degrees of freedom in the system. This method takes a selected group of representative atoms in the atomic microstructure as a bead and uses the proportional relationship of atomic scale atomic potentials in MD simulation to define the interaction between beads and solve the motion equation of beads. This article proposed a coarse-grained potential for graphite based on the modification of Airebo potential, with an n3: 1 mapping, and established a corresponding n3: 1 coarse-grained model, where one bead represents the number of n3 atoms. The results indicated that the coarse-grained model well reflected the basic thermal and mechanical properties of graphite. In addition, this article also proposed a coarsening method for the Lennard–Jones (L–J) potential function parameters and established a coarse-grained wetting model with n = 2. The results indicate that the coarse-grained wetting model can effectively predict the wetting performance of copper droplets on graphite in only 60% of the time using the all-atom model. The coarsening potential function and model proposed in this article are also applicable to graphite with rough surfaces. It can be anticipated that coarse-grained molecular dynamics methods will have more applications in the future, as they can handle large-scale calculations more quickly.

Enhanced particle removal ability of two representative nonionic surfactants: A reasonable interpretation based on DFT and coarse-grained molecular dynamics methods

Cleaning of copper surfaces after chemical mechanical polishing is indispensable for removing particle residues, which is a critical step in integrated circuits manufacturing. This paper introduces two nonionic surfactants, coconut oil diethanolamine (CDEA) and fatty alcohol polyoxyethylene ether (AEO-9), into a basic alkaline cleaning solution comprising tetraethyl ammonium hydroxide and lysine to investigate their abilities to enhance particle removal performance. Cleaning performance was measured with scanning electron microscopy, which showed that the two surfactants achieved excellent particle removal efficiency (above 95 %), with CDEA performing slightly better in all tested concentrations. Concerning critical micelle concentration, surface tension tests showed that CDEA (at 1500 ppm) had superior wettability and dispersive capacity compared with AEO-9 (at 2000 ppm) and could prevent particle redeposition more effectively. Using density functional theory, CDEA was found to have more polar groups to form hydrogen bonds with the hydrophilic surface of colloidal silica particles or bond with the copper substrate, explaining CDEA’s ability to lower the surface tension of the cleaning solution and improve the zeta potential between particles and the copper substrate. Finally, based on the Martini forcefield, simulations of coarse-grained molecular dynamics provided valuable theoretical insights into the different dispersion abilities of the two nonionic surfactants that CDEA had a more significant diffusion coefficient (0.073 Å2/ps) and was able to form micelles more efficiently in solution compared with that of AEO-9 (0.066 Å2/ps).

Martini 3 OliGo̅mers: A Scalable Approach for Multimers and Fibrils in GROMACS.

Martini 3 is a widely used coarse-grained simulation method for large-scale biomolecular simulations. It can be combined with a Go̅ model to realistically describe higher-order protein structures while allowing the folding and unfolding events. However, as of today, this method has largely been used only for individual monomers. In this article, we describe how the Go̅ model can be implemented within the framework of Martini 3 for a multimer system, taking into account both intramolecular and intermolecular interactions in an oligomeric protein system. We demonstrate the method by showing how it can be applied to both structural stability maintenance and assembly/disassembly of protein oligomers, using aquaporin tetramer, insulin dimer, and amyloid-β fibril as examples. We find that addition of intermolecular Go̅ potentials stabilizes the quaternary structure of proteins. The strength of the Go̅ potentials can be tuned so that the internal fluctuations of proteins match the behavior of atomistic simulation models, however, the results also show that the use of too strong intermolecular Go̅ potentials weakens the chemical specificity of oligomerization. The Martini-Go̅ model presented here enables the use of Go̅ potentials in oligomeric molecular systems in a computationally efficient and parallelizable manner, especially in the case of homopolymers, where the number of identical protein monomers is high. This paves the way for coarse-grained simulations of large protein complexes, such as viral protein capsids and prion fibrils, in complex biological environments.

Fault Diagnosis Method for Wind Turbine Gearbox Based on Ensemble-Refined Composite Multiscale Fluctuation-Based Reverse Dispersion Entropy.

The diagnosis of faults in wind turbine gearboxes based on signal processing represents a significant area of research within the field of wind power generation. This paper presents an intelligent fault diagnosis method based on ensemble-refined composite multiscale fluctuation-based reverse dispersion entropy (ERCMFRDE) for a wind turbine gearbox vibration signal that is nonstationary and nonlinear and for noise problems. Firstly, improved complete ensemble empirical mode decomposition with adaptive noise (ICEEMDAN) and stationary wavelet transform (SWT) are adopted for signal decomposition, noise reduction, and restructuring of gearbox signals. Secondly, we extend the single coarse-graining processing method of refined composite multiscale fluctuation-based reverse dispersion entropy (RCMFRDE) to the multiorder moment coarse-grained processing method, extracting mixed fault feature sets for denoised signals. Finally, the diagnostic results are obtained based on the least squares support vector machine (LSSVM). The dataset collected during the gearbox fault simulation on the experimental platform is employed as the research object, and the experiments are conducted using the method proposed in this paper. The experimental results demonstrate that the proposed method is an effective and reliable approach for accurately diagnosing gearbox faults, exhibiting high diagnostic accuracy and a robust performance.

Dissipative Particle Dynamics Simulation of Protein Folding in Explicit and Implicit Solvents: Coarse-Grained Model for Atomic Resolution.

Advancements have been made to dissipative particle dynamics (DPD), a robust coarse-grained (CG) simulation method, to study the folded structures of four miniproteins (1L2Y, 1WN8, 1YRF, and 2I9M) in explicit and implicit solvents. In this endeavor, we aim to establish model parametrization and enhance computational efficiency. Unlike traditional CG models that use empirical force parameters, ex-force parameters (r0(ex), , δd, δp) of DPD particles constructed for specific research purposes can be obtained from atomistic molecular dynamics simulations. On the other hand, im-force parameters (r0(im), c, σ) can be derived from ex-DPD simulations, according to the underlying thermodynamic theory. Based on a mapping scheme proposed for the modeling of amino acids, all-atom proteins can be converted into a CG model. Both ex-/im-DPDs are then carried out to investigate the folding pathways of the four mini-proteins. Structural analysis of the RMSDs shows that the im-simulated proteins have greater structural similarity to native proteins than the ex-simulated ones. The constructed CG models achieve a resolution of Angstrom (Å), a level normally associated with atomic models. Additionally, speed tests reveal that im-DPD accelerates the simulation process and significantly improves simulation efficiency.

Uncovering nonequilibrium from unresolved events.

Closely related to the laws of thermodynamics, the detection and quantification of disequilibria are crucial in unraveling the complexities of nature, particularly those beneath observable layers. Theoretical developments in nonequilibrium thermodynamics employ coarse-graining methods to consider a diversity of partial information scenarios that mimic experimental limitations, allowing the inference of properties such as the entropy production rate. A ubiquitous but rather unexplored scenario involves observing events that can possibly arise from many transitions in the underlying Markov process-which we dub multifilar events-as in the cases of exchanges measured at particle reservoirs, hidden Markov models, mixed chemical and mechanical transformations in biological function, composite systems, and more. We relax one of the main assumptions in a previously developed framework, based on first-passage problems, to assess the non-Markovian statistics of multifilar events. By using the asymmetry of event distributions and their waiting times, we put forward model-free tools to detect nonequilibrium behavior and estimate entropy production, while discussing their suitability for different classes of systems and regimes where they provide no new information, evidence of nonequilibrium, a lower bound for entropy production, or even its exact value. The results are illustrated in reference models through analytics and numerics.

Numerical investigation of particle behavior and erosion wear in a centrifugal pump using coarse-grained Discrete Element Method

Erosion wear presents a significant challenge in the operation of hydrodynamic systems, particularly in centrifugal pumps. To address this issue, this study employs the coarse-grained Discrete Element Method (CGDEM) coupled with Computational Fluid Dynamics (CFD) to predict erosion wear areas and elucidate particle behavior dynamics. The simulation considers spherical solid particles (brown corundum) with diameters ranging from 0.3 mm to 1 mm and volume fractions of 15% and 5%. The flow field is calculated using the Navier-Stokes equations and the shear-stress-transport (SST) k-ω model. Particle movement is tracked using Newton’s second law, considering drag and gravitational forces. Results indicate that the highest turbulent kinetic energy (TKE) occurs at a flow rate of Q = 15 m3/h. Moreover, the pressure reduces slightly when the flow rate increases, although flow velocity might increase with volumetric discharge. Additionally, increased particle diameter leads to greater erosion wear, particularly on blade leading edges, the volute tongue area, and the volute casing belly region. As particle mass concentration intensifies, wear rates on the volute wall and inner hub edge also increase. The study’s robustness is demonstrated through alignment between experimental and numerical data. This investigation offers valuable insights into erosion dynamics in centrifugal pumps, crucial for enhancing operational efficiency and mitigating erosion-induced challenges in these critical systems.

A fast local search algorithm for minimum sum coloring problem on massive graphs

The minimum sum coloring problem (MSCP) is an important extension of the graph coloring problem with wide real-world applications. Compared to the classic graph coloring problem, where lots of methods have been developed and even massive graphs with millions of vertices can be solved well, few works have been done for the MSCP, and no specialized MSCP algorithms are available for solving massive graphs. This paper explores how to solve MSCP on massive graphs, and then proposes a fast local search algorithm for the MSCP based on three main ideas including a coarse-grained reduction method, two kinds of scoring functions and selection rules as well as a novel local search framework. Experiments are conducted to compare our algorithm with several state-of-the-art algorithms on massive graphs. The proposed algorithm outperforms previous algorithms in almost all the massive graphs and also improves the best-known solutions for some conventional instances, which demonstrates the performance and robustness of the proposed algorithm.

Coarse-Grained Approach to Simulate Signatures of Excitation Energy Transfer in Two-Dimensional Electronic Spectroscopy of Large Molecular Systems.

Two-dimensional electronic spectroscopy (2DES) has proven to be a highly effective technique in studying the properties of excited states and the process of excitation energy transfer in complex molecular assemblies, particularly in biological light-harvesting systems. However, the accurate simulation of 2DES for large systems still poses a challenge because of the heavy computational demands it entails. In an effort to overcome this limitation, we devised a coarse-grained 2DES method. This method encompasses the treatment of the entire system by dividing it into distinct weakly coupled segments, which are assumed to communicate predominantly through incoherent exciton transfer. We first demonstrate the efficiency of this method through simulation on a model dimer system, which demonstrates a marked improvement in calculation efficiency, with results that exhibit good concordance with reference spectra calculated with less approximate methods. Additionally, the application of this method to the light-harvesting antenna 2 (LH2) complex of purple bacteria showcases its advantages, accuracy, and limitations. Furthermore, simulating the anisotropy decay in LH2 induced by energy transfer and its comparison with experiments confirm that the method is capable of accurately describing dynamical processes in a biologically relevant system. This method presented lends itself to an extension that accounts for the effect of intrasegment relaxation processes on the 2DES spectra, which for computational efficiency are ignored in the implementation reported here. It is envisioned that the method will be employed in the future to accurately and efficiently calculate 2D spectra of more extensive systems, such as photosynthetic supercomplexes.

An Exact Theory of Causal Emergence for Linear Stochastic Iteration Systems.

After coarse-graining a complex system, the dynamics of its macro-state may exhibit more pronounced causal effects than those of its micro-state. This phenomenon, known as causal emergence, is quantified by the indicator of effective information. However, two challenges confront this theory: the absence of well-developed frameworks in continuous stochastic dynamical systems and the reliance on coarse-graining methodologies. In this study, we introduce an exact theoretic framework for causal emergence within linear stochastic iteration systems featuring continuous state spaces and Gaussian noise. Building upon this foundation, we derive an analytical expression for effective information across general dynamics and identify optimal linear coarse-graining strategies that maximize the degree of causal emergence when the dimension averaged uncertainty eliminated by coarse-graining has an upper bound. Our investigation reveals that the maximal causal emergence and the optimal coarse-graining methods are primarily determined by the principal eigenvalues and eigenvectors of the dynamic system's parameter matrix, with the latter not being unique. To validate our propositions, we apply our analytical models to three simplified physical systems, comparing the outcomes with numerical simulations, and consistently achieve congruent results.

A coarse and fine-grained deep multi view subspace clustering method for unsupervised fault diagnosis of rolling bearings

Abstract To address the issue of insufficient characterization of fault features in inherent vibration data that affects the performance of unsupervised learning-based fault diagnosis, a coarse and fine-grained deep multi view subspace clustering method (CFG-DMVSC) for unsupervised fault diagnosis of rolling bearings is proposed. The proposed method designs a convolutional autoencoder network based on the Gramian angular field transformation for multi-signal analysis domains. A multi-view coarse-grained self-expressive method based on information entropy is designed to handle differences in information across different views. Furthermore, a fine-grained common and independent information separation loss function based on mutual information is proposed to ensure compactness among multiple views. Both the Case Western Reserve University rolling bearing dataset and privately built bearing fault test bench data demonstrate that, compared to existing methods, the proposed method can perform coarse and fine-grained division in multi-view subspaces, achieving better clustering diagnosis performance on the extracted common information among views.

A fast dual-module hybrid high-dimensional feature selection algorithm

When dealing with large-scale datasets, high-dimensional feature selection plays a crucial role in improving the performance and interpretability of machine learning models. However, it still faces the problems of the “dimensionality curse” and high computational cost in dealing with high-dimensional datasets. In this paper, we develop a dual-module hybrid feature selection algorithm based on correlation filtering and a multi-objective evolutionary algorithm based on dynamic variation distance (HMOFS-CFDVD), which aims to reduce the computational cost of high-dimensional features and search space. Firstly, a coarse-grained feature filtering method based on correlation is proposed, enabling the algorithm to quickly identify potentially better feature subsets in the later stages. After that, a fine-grained feature selection is further implemented based on the multi-objective evolutionary algorithm with sample variation distance to obtain a high-quality feature subset. This stage incorporates a novel population initialization method with self-regulation, an adaptive evolution strategy, an optimal sample selection mechanism based on dynamic sample distance, and an improved diversity mechanism to enhance the evolutionary performance. Moreover, the feature relevance metric is introduced as a third objective to improve the overall algorithm's performance. Experimental results on 12 high-dimensional feature datasets demonstrate that the proposed HMOFS-CFDVD achieves high accuracy and produces a smaller subset of features.

Intelligent Fault Diagnosis Method for Transformer Driven by Multiple Vibration Data

Abstract Early monitoring and early warning of transformer failure are crucial to ensure the safe and dependable functioning of the power system, as transformers are essential equipment within the system. On multivariate vibration data, this work presents a sophisticated approach for diagnosing faults in transformers which integrates entropy model and machine learning. Firstly, a new coarse-grained method is introduced into the improved multiscale fuzzy entropy to improve the disadvantage of the traditional entropy model with large entropy fluctuation under high-scale factors. Furthermore, the features of transformer multivariate vibration signals are extracted by the entropy model. Finally, the extreme learning machine is utilized to achieve efficient detection of transformer defects. This method provides a useful method reference for fault diagnosis of power equipment with multi-vibration signal, and has good reference value.

Influence of cyclone combustor structure on the motion behaviors of solid fuel particles

A coarse-grained computational fluid dynamics-discrete element method (CFD-DEM) was established to investigate the influence of cyclone combustor structure on the motion behaviors of solid fuel particles. Four structural parameters, i.e., primary inlet number (Pn), secondary inlet number (Sn), secondary inlet location (SL), and guide blade angle (α), were examined. Compared to the base structure with a single primary inlet, the addition of extra primary inlet, secondary inlet, or diversion structure elevates the global parcel location. Increasing Pn and Sn decreases the average parcel residence time, while augmenting SL and α generally lengthens it. Meanwhile, increasing Pn, Sn, and SL or decreasing α broadly reduces the wall parcel-hold ratio. Furthermore, the inclusion of secondary inlet and diversion structure increases the probability of small parcels occurring in the near-wall zone. These observations are helpful for optimizing the design and operation of cyclone combustors.

Validation on fly ash DEM microparameters based on coarse-grained method

The computational efficiency of the discrete element simulation process is often a limiting factor. However, the use of the coarse-grained method, which employs dimensional analysis to amplify particles, significantly improves computational efficiency. This research utilizes the coarse-grained method to amplify particles and examines the impact of linear models on the mechanical of fly ash. By examining parameters such as magnification, stiffness, sliding friction coefficient, and rolling friction coefficient, we establish the relationship between fly ash cohesion, internal friction angle, and microscopic parameters in discrete element method (DEM) simulation. The simulation results demonstrate that the shear strength of fly ash increases as magnification, stiffness, and sliding friction coefficient increase. The rolling friction coefficient, on the other hand, has minimal effect on shear strength. The elastic potential energy of the system is unaffected by magnification, with stiffness inversely proportional to adaptable potential power. However, sliding and rolling friction coefficients positively impact adaptable potential power. Furthermore, this study proposes a practical simplification method for amplification of materials with small particle sizes. The comparison between numerical simulation and experimental data yields consistent results. As a result, this study provides a foundation for the application of large-scale fly ash transportation, storage, and other engineering purposes.