Multiscale and Auto-Tuned Semi-Supervised Deep Subspace Clustering and Its Application in Brain Tumor Clustering

In this paper,we introduce a novel Multi-scale and Auto-tuned Semi-supervised Deep Subspace Clustering(MAS-DSC)algorithm,aimed at addressing the challenges of deep subspace clustering in high-dimensional real-world data,particularly in the field of medical imaging.Traditional deep subspace clustering algorithms,which are mostly unsupervised,are limited in their ability to effectively utilize the inherent prior knowledge in medical images.Our MAS-DSC algorithm incorporates a semi-supervised learning framework that uses a small amount of labeled data to guide the clustering process,thereby enhancing the discriminative power of the feature representations.Additionally,the multi-scale feature extraction mechanism is designed to adapt to the complexity of medical imaging data,resulting in more accurate clustering performance.To address the difficulty of hyperparameter selection in deep subspace clustering,this paper employs a Bayesian optimization algorithm for adaptive tuning of hyperparameters related to subspace clustering,prior knowledge constraints,and model loss weights.Extensive experiments on standard clustering datasets,including ORL,Coil20,and Coil100,validate the effectiveness of the MAS-DSC algorithm.The results show that with its multi-scale network structure and Bayesian hyperparameter optimization,MAS-DSC achieves excellent clustering results on these datasets.Furthermore,tests on a brain tumor dataset demonstrate the robustness of the algorithm and its ability to leverage prior knowledge for efficient feature extraction and enhanced clustering performance within a semi-supervised learning framework.

Prediction of multiscale laminae structure and reservoir quality in fine-grained sedimentary rocks:The Permian Lucaogou Formation in Jimusar Sag,Junggar Basin

Fine-grained sedimentary rocks have become a research focus as important reservoirs and source rocks for tight and shale oil and gas.Laminae development determines the accumulation and production of tight and shale oil and gas in fine-grained rocks.However,due to the resolution limit of conventional logs,it is challenging to recognize the features of centimeter-scale laminae.To close this gap,complementary studies,including core observation,thin section,X-ray diffraction(XRD),conventional log analysis,and slabs of image logs,were conducted to unravel the centimeter-scale laminae.The laminae recognition models were built using well logs.The fine-grained rocks can be divided into laminated rocks(lamina thickness of<0.01 m),layered rocks(0.01-0.1 m),and massive rocks(no layer or layer spacing of>0.1 m)according to the laminae scale from core observations.According to the mineral superposition assemblages from thin-section observations,the laminated rocks can be further divided into binary,ternary,and multiple structures.The typical mineral components,slabs,and T2spectrum distributions of various lamina types are unraveled.The core can identify the centimeter-millimeter-scale laminae,and the thin section can identify the millimeter-micrometer-scale laminae.Furthermore,they can detect mineral types and their superposition sequence.Conventional logs can identify the meter-scale layers,whereas image logs and related slabs can identify the laminae variations at millimeter-centimeter scales.Therefore,the slab of image logs combined with thin sections can identify laminae assemblage characteristics,including the thickness and vertical assemblage.The identification and classification of lamina structure of various scales on a single well can be predicted using conventional logs,image logs,and slabs combined with thin sections.The layered rocks have better reservoir quality and oil-bearing potential than the massive and laminated rocks.The laminated rocks’binary lamina is better than the ternary and multiple layers due to the high content of felsic minerals.The abovementioned results build the prediction model for multiscale laminae structure using well logs,helping sweet spots prediction in the Permian Lucaogou Formation in the Jimusar Sag and fine-grained sedimentary rocks worldwide.

Study on the evolution of solid–liquid–gas in multi-scale pore methane in tectonic coal

The rich accumulation of methane(CH_(4))in tectonic coal layers poses a significant obstacle to the safe and efficient extraction of coal seams and coalbed methane.Tectonic coal samples from three geologically complex regions were selected,and the main results obtained by using a variety of research tools,such as physical tests,theoretical analyses,and numerical simulations,are as follows:22.4–62.5 nm is the joint segment of pore volume,and 26.7–100.7 nm is the joint segment of pore specific surface area.In the dynamic gas production process of tectonic coal pore structure,the adsorption method of methane molecules is“solid–liquid adsorption is the mainstay,and solid–gas adsorption coexists”.Methane stored in micropores with a pore size smaller than the jointed range is defined as solid-state pores.Pores within the jointed range,which transition from micropore filling to surface adsorption,are defined as gaseous pores.Pores outside the jointed range,where solid–liquid adsorption occurs,are defined as liquid pores.The evolution of pore structure affects the methane adsorption mode,which provides basic theoretical guidance for the development of coal seam resources.

A multiscale differential-algebraic neural network-based method for learning dynamical systems

The objective of dynamical system learning tasks is to forecast the future behavior of a system by leveraging observed data.However,such systems can sometimes exhibit rigidity due to significant variations in component parameters or the presence of slow and fast variables,leading to challenges in learning.To overcome this limitation,we propose a multiscale differential-algebraic neural network(MDANN)method that utilizes Lagrangian mechanics and incorporates multiscale information for dynamical system learning.The MDANN method consists of two main components:the Lagrangian mechanics module and the multiscale module.The Lagrangian mechanics module embeds the system in Cartesian coordinates,adopts a differential-algebraic equation format,and uses Lagrange multipliers to impose constraints explicitly,simplifying the learning problem.The multiscale module converts high-frequency components into low-frequency components using radial scaling to learn subprocesses with large differences in velocity.Experimental results demonstrate that the proposed MDANN method effectively improves the learning of dynamical systems under rigid conditions.

Simultaneous enhancements of strength and toughness by multiscale lamellar structure in Ti_(2)AlNb based intermetallic

Multi-scale lamellar structure significantly improves toughness of Ti_(2)AlNb based alloys,which are inher-ently brittle intermetallics,without compromising their strength.This structure was achieved through-B2-transus-forging(TBTF)combined with O+B2 two-phase region heat treatments.Various types of multi-scale lamellar structures were obtained by controlling the cooling rate after TBTF.These variations were mainly attributed to differences in the distribution,content,and size of the thick lamellar O phase and the size and crystallographic orientation of B2 grain.By analyzing the microstructural characteristics and crystallographic orientation near the crack propagation path,it was found that the crack propaga-tion resistance of thick lamellae,sub grain and grain boundaries(GBs)O phase increased sequentially,accompanied by more tortuous crack propagation path.Moreover,B2 grains with high misorientation significantly deflected the crack propagation by cleavage ridges between adjoining cleavage planes.Addi-tionally,the development of numerous secondary cleavage ridges,resulting from the transition through varying secondary cleavage planes in distinct sub B2 grains,further hindered the quick propagation of cracks.It was clarified that the cleavage planes were dominantly belonging to{110}.These findings pro-vided valuable guidance for the design of damage tolerance strategies for Ti_(2)AlNb-based intermetallics.

Multiscale patterning of a metallic glass using sacrificial imprint lithography

Bulk metallic glasses(BMGs)have been developed as a means to achieve durable multiscale,nanotextured surfaces with desirable properties dictated by topography for a multitude of applications.One barrier to this achievement is the lack of a bridging technique between macroscale thermoplastic forming and nanoimprint lithography,which arises from the difficulty and cost of generating controlled nanostructures on complex geometries using conventional top-down approaches.This difficulty is compounded by the necessary destruction of any resulting reentrant structures during rigid demolding.We have developed a generalized method to overcome this limitation by sacrificial template imprinting using zinc oxide(ZnO)nanostructures.It is established that such structures can be grown inexpensively and quickly with tunable morphologies on a wide variety of substrates out of solution,which we exploit to generate the nanoscale portion of the multiscale pattern through this bottom-up approach.In this way,we achieve metallic structures that simultaneously demonstrate features from the macroscale down to the nanoscale,requiring only the top-down fabrication of macro/microstructured molds.Upon detachment of the formed part from the multiscale molds,the ZnO remains embedded in the surface and can be removed by etching in mild conditions to both regenerate the mold and render the surface of the BMGs nanoporous.The ability to pattern metallic surfaces in a single step on length scales from centimeters down to nanometers is a critical step toward fabricating devices with complex shapes that rely on multiscale topography for their intended functions,such as biomedical and electrochemical applications.

A novel active-passive protection strategy endows carbonyl iron with excellent intelligent anti-corrosion and microwave absorption dual functional characteristics

Carbonyl iron absorbers(CI)face significant challenges in practical applications,such as corrosion,interface bonding failure,detachment,and high maintenance costs.Herein,we have developed intelligent self-healing technology based on proactive/passive mechanisms via in situ synthesis of self-healing factors(polydopamine/benzotriazole(PDA/BTA))and physical barrier layers(SiO_(2)/1,1,1,3,3,3-hexamethyl disilazane,SiO_(2)/HMDS)to enhance corrosion resistance,while also being compatible with efficient microwave absorption characteristics.The unique multiscale structure gives full play to the utilization of the roles of each functional layer,including the intelligent self-healing features of PDA/BTA,physical shielding and spatial confinement characteristics of SiO_(2)/HMDS,and magnetic-dielectric synergistic mechanism resulted from the good impedance matching characteristics,the conduction loss,the interfacial polarization loss and natural resonance.The asfabricated composites achieved an exceptional minimum reflection loss value of-55.4 dB at 10.7 GHz and the effective absorption band of 7.6 GHz.Moreover,it still exhibits obvious self-healing and corrosion resistance characteristics after 360 h corrosion treatment,ascribed to the self-healing mechanism of PDA/BTA and the blocking intervention effect of SiO_(2)/HMDS.This work is considered to pave the way for the synthesis of high-performance magnetic absorbers,especially in enhancing their intelligent self-healing ability in corrosive environments.

Multiscale structural engineering of carbon nitride for enhanced photocatalytic H2O2 production

Carbon nitride(C_(3)N_(4))holds great promise for photocatalytic H_(2)O_(2)production from oxygen reduction.In spite of great research efforts,they still suffer from low catalytic efficiency primarily limited by the fast recombination of photogenerated charge carriers.In this work,we report the multiscale structural engineering of C_(3)N_(4)to significantly improve its optoelectronic properties and consequently photocatalytic performance.The product consists of porous spheres with high surface areas,abundant nitrogen defects,and alkali metal doping.Under visible light irradiation,our catalyst shows a remarkable H_(2)O_(2)production rate of 3,080μmol·g^(−1)·h^(−1),which is more than 10 times higher than that of bulk C_(3)N_(4)and exceeds those of most other C_(3)N_(4)-based photocatalysts.Moreover,the catalyst exhibits great stability,and can continuously work for 15 h without obvious activity decay under visible light irradiation,eventually giving rise to a high H_(2)O_(2)concentration of ca.45 mM.

Multiscale cellulose-based fireproof and thermal insulation gel materials with water-regulated forms

Different forms of construction materials(e.g.,paints,foams,and boards)dramatically improve the quality of life.With the increasing environmental requirements for buildings,it is necessary to develop a comprehensive sustainable construction material that is flexible in application and exhibits excellent performance,such as fireproofing and thermal insulation.Herein,an adjustable multiform material strategy by water regulation is proposed to meet the needs of comprehensive applications and reduce environmental costs.Multiform gels are constructed based on multiscale cellulose fibers and hollow glass microspheres,with fireproofing and thermal insulation.Unlike traditional materials,this multiscale cellulose-based gel can change forms from dispersion to paste to dough by adjusting its water content,which can realize various construction forms,including paints,foams,and low-density boards according to different scenarios and corresponding needs.

Assembling 2D Ni-Co Nanosheets onto Mo_(2)C Nanorod towards Efficient Electrocatalytic Hydrogen Evolution

A novel electrocatalyst,Ni-Co/β-Mo_(2)C@C,was rationally designed to enhance the efficiency of the hydrogen evolution reaction(HER)in this work.Assembled with two-dimensional Ni-Co nanosheets onto Mo_(2)C nanorods coated with a thin carbon shell,the catalyst demonstrates remarkable performance,including low overpotential(η_(10)=57 mV)and reduced Tafel slope(63 mV·dec^(–1))in 0.5 mol·L^(–1)H_(2)SO_(4) electrolyte.This innovative design strategy provides abundant active sites and efficient electron/ion transport pathways,effectively shortening reactant diffusion distances and enhancing electrocatalytic activity.Additionally,the carbon shell coating protects the catalyst from etching and agglomeration,ensuring its durability.This work presents a promising approach for engineering highly efficient metal carbide-based HER catalysts through tailored composition and nanostructure design.

Pt single atoms coupled with Ru nanoclusters enable robust hydrogen oxidation for high-performance anion exchange membrane fuel cells

The sluggish reaction kinetics of alkaline hydrogen oxidation reaction(HOR)is one of the key challenges for anion exchange membrane fuel cells(AEMFCs).To achieve robust alkaline HOR with minimized cost,we developed a single atom-cluster multiscale structure with isolated Pt single atoms anchored on Ru nanoclusters supported on nitrogen-doped carbon nanosheets(Pt1-Ru/NC).The well-defined structure not only provides multiple sites with varied affinity with the intermediates but also enables simultaneous modulation of different sites via interfacial interaction.In addition to weakening Ru–H bond strength,the isolated Pt sites are heavily involved in hydrogen adsorption and synergistically accelerate the Volmer step with the help of Ru sites.Furthermore,this catalyst configuration inhibits the excessive occupancy of oxygen-containing species on Ru sites and facilitates the HOR at elevated potentials.The Pt1-Ru/NC catalyst exhibits superior alkaline HOR performance with extremely high activity and excellent CO-tolerance.An AEMFC with a 0.1 mg·cmPGM^(−2)loading of Pt1-Ru/NC anode catalyst achieves a peak powder density of 1172 mW·cm^(−2),which is 2.17 and 1.55 times higher than that of Pt/C and PtRu/C,respectively.This work provides a new catalyst concept to address the sluggish kinetics of electrocatalytic reactions containing multiple intermediates and elemental steps.

A dual-eddy EMMS-based turbulence model for laminar-turbulent transition prediction

Turbulence is a century-old physics problem,and the prediction of laminar-turbulent transition remains a major challenge in computational fluid dynamics(CFD).This paper proposes a new conceptual multiscale-structure flow system consisting of a nonturbulent part and two types of turbulent eddies with different properties.The stability criterion for turbulent transition flows,based on the principle of compromise-in-competition between viscosity and inertia,is used to obtain model closure.The multiscale-structure concept and stability criterion are the characteristics of the dual-eddy energy-minimization multiscale(EMMS)-based turbulence model.The solved heterogeneous structure parameters and energy dissipation rate are analyzed,which reveal the laminar-turbulent transition process.To validate the dual-eddy EMMS-based turbulence model,three benchmark problems,namely,the transitional flows over the flat plate boundary layer with zero pressure gradient,NACA0012,and Aerospatiale-A airfoils,were simulated.The simulation was performed by combining the optimized results from the proposed model with the equations of the well-known κ-ω shear stress transfer(SST)turbulence model.The numerical results show that the dual-eddy EMMS-based turbulence model improves the prediction in the laminar-turbulent transition process.This demonstrates the soundness of using the multiscale-structure concept in turbulent flows to establish the turbulence transition model by considering the principle of compromise-in-competition between viscosity and inertia.

C/Cu多孔微球固载Co3O4纳米片作为锂离子电池的高性能负极材料

过渡金属氧化物(TMOs)用作电极材料时,会在循环过程中产生严重的体积变化,并且其自身的导电率也较低,因此它的电化学性能较差.设计和开发独特的TMOs纳米结构并将其与导电碳基底相结合是改善其电化学性能的有效策略.本工作中,我们设计了一种C/Cu多孔微球,并通过原位合成在碳壁上垂直生长Co_(3)O_(4)纳米片.作为导电基底,C/Cu多孔微球提供了多尺度孔隙网络和大的电极/电解质接触界面,显著改善了电子和离子扩散动力学.原位合成的Co_(3)O_(4)纳米片牢牢地固定在碳壁上,从而提高了复合微球在长期循环中的结构稳定性.得益于独特的结构特征,用作锂离子电池负极材料的C/Cu@Co_(3)O_(4)复合多孔微球表现出优异的倍率性能、高充电比容量和出色的循环稳定性.