Unified Relationship Research Articles

Learning Double-Level Relationship Networks for image captioning

Image captioning aims to generate descriptive sentences to describe image main contents. Existing attention-based approaches mainly focus on the salient visual features in the image. However, ignoring the learning relationship between local features and global features may cause local features to lose the interaction with global concepts, generating impropriate or inaccurate relationship words/phrases in the sentences. To alleviate the above issue, in this work we propose the Double-Level Relationship Networks (DLRN) that novelly exploits the complementary local features and global features in the image, and enhances the relationship between features. Technically, DLRN builds two types of networks, separate relationship network and unified relationship embedding network. The former learns different hierarchies of visual relationship by performing graph attention for local-level relationship enhancement and pixel-level relationship enhancement respectively. The latter takes the global features as the guide to learn the local–global relationship between local regions and global concepts, and obtains the feature representation containing rich relationship information. Further, we devise an attention-based feature fusion module to fully utilize the contribution of different modalities. It effectively fuses the previously obtained relationship features and original region features. Extensive experiments on three typical datasets verify that our DLRN significantly outperforms several state-of-the-art baselines. More remarkably, DLRN achieves the competitive performance while maintaining notable model efficiency. The source code is available at the GitHub https://github.com/RunCode90/ImageCaptioning.

Genetic algorithm-assisted data-driven model for boundary shear distribution and stage-discharge: Compound open channel flows

In this paper, a unified method has been proposed for the percentage shear force carried by floodplains using fewer non-dimensional parameters such as the floodplain's percentage area, the ratio of Manning’s roughness and the depth ratio. This new data-driven dynamic model for percentage shear force is obtained using a genetic algorithm (GA) program, a well-documented machine-learning software, which can examine the existing relationship among the variables and explore the influencing factors. GA facilitated a unified relationship between apparent shear force and the chosen parameters. The new proposed model is simple and accurate compared to the previous models, which were either complex or distinct for different configurations. The efficiency of the new model shows that the most predicted test case results are under the 5% error cap. The cohesive expression derived for smooth and roughened compound channels is the most significant advantage of the GA-based data-driven model, providing a simple and easy-use formula for engineers to apply for broad applications. Compared with the other available approaches, the model proposed provides the most accurate discharge prediction for various unique datasets.

Energy-based assessment of cyclic liquefaction behavior of clean and silty sand under sustained initial stress conditions

In practical engineering, sand deposits containing some finer particles usually sustain an initial static shear stress. This study investigates the sustained initial stress effect on liquefaction response of saturated sand through conducting cyclic undrained triaxial tests. The obtained response of sand samples containing a small amount of silty fines is compared to that of clean sand under similar relative densities. The results exhibit that various stress conditions result in two response patterns named as cyclic mobility and residual deformation accumulation. It is shown that the addition of silty fines can either enhance or reduce the cyclic liquefaction resistance of sand, depending on the fines content, while the sustained initial stress plays a beneficial role in resisting cyclic failure for samples with or without the addition of fines under otherwise identical testing conditions. It is also found that the dissipated energy accumulated in both the silty and clean sand follows a power-law relationship to the generated residual pore pressure during cyclic loading. A unified relationship is further established between the liquefaction resistance and dissipated energy, which may be useful for developing an energy-based assessment of in-situ soils using strength data from laboratory experiments.

A unified constitutive relationship with the internal state variables of grain evolution behavior and the application in numerical simulation of cross wedge rolling

Controlling the grain size is an essential issue in the metal plastic working processes. Based on the consensus view that grain boundaries are impassable borders for dislocations and that the number of dislocations within a grain affects the stress builds up among the adjacent grains, altering grain size has a great influence on the post-forming mechanical properties of the deformed workpieces. In the present research, the deformation behaviors of 20CrMoH, a typical hardenability steel, are investigated and the microstructure evolution of grain growth experiencing heat load and grain refinement under deformation load is also uncovered. According to the experimental stress-strain curves and grain size variation data, a set of internal variable constitutive equations are established, in which the material constants are calibrated by a genetic optimization algorithm. Comparing the predicted curves with the experimental values, it can be concluded that the constitutive equations can accurately reproduce the thermal deformation behaviors and grain size variation of the steel. The correlation coefficient of the calculated and experimental values is 0.9861, and the average relative error approximates 4%. The unified constitutive equations are compiled into a commercial finite element software with the help of user-defined subroutine interfaces. The simulation results demonstrate that the maximum accuracy error of a hollow shaft size is about 10%, and the error in the average grain size is 12.1%. It is confirmed that the internal variable constitutive equations describing the evolution of grain size established in this paper are suitable for the numerical simulation of the bulk-forming process.

Stiffening and softening in the power-law rheological behaviors of cells

Living cells are a complex soft material with fascinating mechanical properties. Confusingly, experiments have shown that cells exhibit stiffening and more solid-like behaviors under uniaxial stretches or shear, while they present softening and more fluid-like behaviors under biaxial stretches. For both of these seemingly paradoxical stiffening and softening rheological behaviors, cells often exhibit a robust power-law rheological characteristic. Here, based on the structural features, we propose a cellular structural model to investigate these rheological behaviors of cells under different loading conditions. It is found that this structural model can naturally capture the stiffening and softening behaviors in the power-law rheological responses of cells, depending on the loading conditions. Both stiffening and softening of cells originate from changes in the configuration of the discrete cytoskeleton: stiffening from the rotation of the microtubules to the loading direction and softening from the elastic buckling of individual microtubules. Moreover, for both stiffening and softening in the rheological behaviors of cells, there exists a unified relationship that the power-law exponent decreases linearly with the cellular stiffness in a semi-logarithmic coordination. We further present that a self-similar hierarchical model can be used to analyze this unified relationship. This study not only provides a discrete cellular structural model to capture the essential mechanisms of cellular rheology, but also suggests that the scaling rheological exponent may be treated as a mechanical marker for monitoring cellular healthy states.

Crossover from diffusive to convective regimes during miscible displacements in 2D porous media

Miscible displacements are involved in many applications, including enhanced oil recovery and groundwater remediation. When a less viscous and miscible fluid displaces another more viscous one, an unstable front occurs, and its evolution is controlled by the interplay between diffusion and convection. Much of the research has focused on the instability of the mixing zone front, but the crossover of flow regimes during miscible flooding remains unclear. Here, we conduct unstable miscible displacement experiments in rectangular 2D microfluidic models with controlled pore-scale disorder under various flow rates. By quantifying the features of mixing zone and sweep efficiency, we observe and characterize the crossover from diffusive regime to convective regime during flooding. We further study the effects of the flow rate and pore-scale disorder on the crossover time. We show that the increase in flow rate results in an earlier crossover by promoting the convection in the horizontal direction. Moreover, the increasing pore-scale disorder also results in an earlier crossover by inducing a more unstable mixing zone front. We then perform scaling analysis of the crossover time by introducing the coefficient of variation that represents the heterogeneity of the flow geometry. We establish a unified relationship between the scaled crossover time and the Péclet number, which provides a useful approach to predict the crossover time under various conditions. Our work reveals the crossover between flow regimes and improves our understanding of the front evolution during unstable miscible displacement.

Efficient image-based soiling quantification using histograms of a checkered pattern

In this study, a model correction and a new brightness ratio calculation method were proposed to improve a previously reported checkered pattern image-based soiling quantification technique. The image’s black-white brightness ratio (BWR) was obtained for various levels of soiling and camera altitude angles, using only the histogram of the visible-light image of a checkered pattern, consequently denoted hBWR. Normal-incidence soiling loss of a PV reference cell subjected to the same soiling levels as the checkered pattern was determined using the short-circuit current method and a solar simulator. The hBWR quantity was found to be effectively identical to the BWR quantity determined from a more complex image processing method previously used. Also, a unified relationship was found between hBWR and soiling loss corrected by the sine of the camera altitude angle, for all soiling levels and all camera altitude angles. These findings show the feasibility of implementing the image-based soiling quantification technique using readily available photography equipment and generic image processing tools.

A unified parameter model based on machine learning for describing microbial transport in porous media

The transport and retention of microorganisms are typically described using attachment/detachment and straining/liberation models. However, the parameters in the models varied significantly, posing a significant challenge to describe microbial transport under different environmental conditions. A neural network (ANN) model was developed in this study to link the parameters in the model with the factors influencing microbial transport including the properties of microorganisms such as size and surface potentials, and the properties of porous media such as grain size and porosity, and flow conditions. Exhaustive search of literature renders 420 sets of experimental data of microbial transport, which were fitted using the microbial transport model to obtain model parameters. The model parameters, together with the factors influencing microbial transport, were then used to train an ANN model to search for their relationship. An ANN-based parameter relationship was derived and was then used to simulate microbial transport. The simulated results using the relationship roughly matched with the experimental data under different environmental conditions, indicating that a unified relationship was established between the parameters of the microbial transport model and the factors influencing microbial transport, and that microbial transport can be described using the microbial transport model with the ANN-based unified relationship for model parameters.

Analysis of large deformation of deep-buried brittle rock tunnel in strong tectonic active area based on macro and microcrack evolution

The large deformation is a common engineering problem during tunnel excavation. For brittle rock tunnels, the large deformation is closely related to the rock mass structure. In this study, taking the Guanshan tunnel as an example, large deformation of a deep-buried brittle rock tunnel in a strong tectonic active area was analyzed. The large deformation was explained in terms of the surrounding rock (diorite) mechanical properties and the development and evolution of macro and microcracks. The results showed that the diorite in the Guanshan mountain is characterized by high strength and brittleness. However, numerous microcracks are developed in the diorite, which can easily induce a cataclastic brittle failure of the surrounding rock by the transformation of the stress field. In terms of the spatial distribution, the macro and microcracks in the surrounding rock have a good unified relationship, whose development mechanism is related to strong tectonic activity. The stress environment of the surrounding rock changed during the tunnel excavation. A borehole camera revealed that the large deformation of the Guanshan tunnel is synthetically influenced by the macro and microcracks in the surrounding rock. The deformation includes the relaxation of the original macrocracks, fracture of the microcracks, and rock dilatation caused by the opening of the microcracks, among which the fracture of microcracks is the dominance. This study provides a new concept for analyzing large deformation in brittle rock tunnels, indicating that the negative effects on the stability of the surrounding rock induced by the development and evolution of microcracks should be fully considered in strong tectonic active areas.

How Coalescing Bubbles Depart from a Wall.

Bubble evolution plays a fundamental role in boiling and gas-evolving electrochemical systems. One key stage is bubble departure, which is traditionally considered to be buoyancy-driven. However, conventional understanding cannot provide the full physical picture, especially for departure events with small bubble sizes commonly observed in water splitting and high heat flux boiling experiments. Here, we report a new regime of bubble departure owing to the coalescence of two bubbles, where the departure diameter can be much smaller than the conventional buoyancy limit. We show the significant reduction of the bubble base area due to the dynamics of the three-phase contact line during coalescence, which promotes bubble departure. More importantly, combining buoyancy-driven and coalescence-induced bubble departure modes, we demonstrate a unified relationship between the departure diameter and nucleation site density. By elucidating how coalescing bubbles depart from a wall, our work provides design guidelines for energy systems which can largely benefit from efficient bubble departure.

Thermally Decomposed Binary Fe–Cr Alloys: Toward a Quantitative Relationship Between Strength and Structure

Binary Fe–Cr alloys are model alloys for ferritic steels proposed as structural materials for future fusion reactors. They are used to investigate the fundamental mechanisms of their degradation induced by heat and irradiation. Fe–Cr presents a miscibility gap, which induces Cr‐rich (α′) regions in an Fe‐rich (α) matrix. As this causes embrittlement, it is crucial to understand this phase decomposition and its role, starting with the heat impact. Fe–Cr alloys with 5−40 wt% Cr were annealed at 500 °C for up to 1008 h. The microstructure was probed by chemical mapping using scanning transmission electron microscopy with energy‐dispersive X‐ray spectroscopy (EDS) and atom probe tomography, and hardness was assessed by Vickers testing. Increasing Cr content increases hardness, and beyond 15 wt% Cr it further increases upon annealing. At 20 wt% Cr, nanoscale globular α′ precipitates appear, while at 40 wt% Cr an α′‐percolating structure develops. In both cases, the α′ core composition reaches slightly more than 80 at% Cr, and hardness doubles. A unified relationship is found between the alloy strength and the α′ structure and it is shown that this type of hardening is a general mechanism for mature systems, independent of the nominal alloy composition.

Single-step genomic BLUP enables joint analysis of disconnected breeding programs: an example with Eucalyptus globulus Labill.

Single-step GBLUP (HBLUP) efficiently combines genomic, pedigree, and phenotypic information for holistic genetic analyses of disjunct breeding populations. We combined data from two independent multigenerational Eucalyptus globulus breeding populations to provide direct comparisons across the programs and indirect predictions in environments where pedigreed families had not been evaluated. Despite few known pedigree connections between the programs, genomic relationships provided the connectivity required to create a unified relationship matrix, H, which was used to compare pedigree-based and HBLUP models. Stem volume data from 48 sites spread across three regions of southern Australia and wood quality data across 20 sites provided comparisons of model accuracy. Genotyping proved valuable for correcting pedigree errors and HBLUP more precisely defines relationships within and among populations, with relationships among the genotyped individuals used to connect the pedigrees of the two programs. Cryptic relationships among the native range populations provided evidence of population structure and evidence of the origin of landrace populations. HBLUP across programs improved the prediction accuracy of parents and genotyped individuals and enabled breeding value predictions to be directly compared and inferred in regions where little to no testing has been undertaken. The impact of incorporating genetic groups in the estimation of H will further align traditional genetic evaluation pipelines with approaches that incorporate marker-derived relationships into prediction models.

Non-thermal melting of tungsten under intense electronic excitations

Non-thermal effect caused by ultrafast lasers and swift ions in materials are very intriguing, which is of both scientific interest and technological importance. However, the underlying physics of non-thermal effect on ultrafast process remains unclear and the proposed mechanisms have been controversial. Based on the perturbation approximation under tight-binding theory, the non-thermal effect on tungsten (W) are extensively studied. We demonstrate that the non-thermal effect stemmed from the intense electronic excitations induce dramatic decrease in the melting point of W crystal, as well as non-thermal melting inside the W slab. Our analysis shows that the non-thermal forces are essentially responsible for the drop of melting point of the bulk system. Remarkably, the non-thermal effect combined with surface effect on a W film enhance the ordering of the direction of atomic motion near the surface, preventing melting near the surface area, but leading to non-thermal melting in the interior area of the film. Our work also exhibits a unified relationship between the non-thermal melting and the interatomic forces. This relationship is universal in metals and semiconductors irradiated by ultrafast lasers or swift ions, and has been well established long before.

The Undrained Characteristics of Tengger Desert Sand from True Triaxial Testing

Aimed at the characteristics of aeolian sand under rapid construction conditions in desert geotechnical engineering, a series of the true triaxial undrained test were carried out on the GDS apparatus. The 3D deformation, failure, and other characteristics of the dense sand are obtained. Under the condition of same p c , the state transition point where the void water pressure changes from increasing to decreasing appears earlier and leads to enhanced dilatancy with the increase of b, which means the enhanced dilatancy of dense sand caused the increase in strength. The results of the same b shows that the void water pressure generally indicates a decrease at low confining pressure and an increase at high confining pressure, indicating that the aeolian sand shows dilatancy at low confining pressure and contraction at high confining pressure. The state transition point increases with the increase of p c , but all points tend to the same critical state line and state transition line. When b = 0, the critical state line is q = 1.57 p ′ , and the state transition line is q = 1.23 p ′ . When b = 1, the critical state line is q = 1.24 p ′ , and the state transition line is q = 1.04 p ′ . The results at same b obtained the unified critical state line and the state transition line. Therefore, the true triaxial test can obtain the unified relationship of void ratio, p c and b, which overcomes the fact that the existing test cannot consider the influence of b. The test results provide a basis data for the design, construction, and maintenance of geotechnical engineering in Tengger Desert.

Detailed balance limits for inversion in solar-pumped lasers and allied systems.

Solar-pumped lasers and optical amplifiers continue to draw research interest with advances in nanomaterials science and technology. Establishing accurate detailed balance limits for inversion in these systems is essential. In this Letter, we re-examine the threshold limits for inversion in broadband-pumped lasers, with reference to those provided by Roxlo and Yablonvitch [Opt. Lett.8, 271 (1983)OPLEDP0146-959210.1364/OL.8.000271], where they determined the minimum Stokes shift and the minimum ratio of pump band to emission band absorption constants-based on independent consideration of the emission at pump and emission frequencies. In contrast, the derivation here simultaneously accounts for emission in both the pump and emission bands, which in turn leads to a single consolidated inequality that serves to establish the revised threshold requirements for inversion. Upon applying this new unified relationship to solar-pumped devices, a large increase in the minimum required Stokes shifts for 1-sun devices, particularly at larger pump energies and smaller ratios of αp0/αe0, is found. The maximum possible efficiencies of solar-pumped devices are calculated using this new relation.

Understanding Exposure-Receptor Occupancy Relationships for Metabotropic Glutamate Receptor 5 Negative Allosteric Modulators across a Range of Preclinical and Clinical Studies.

The metabotropic glutamate receptor 5 (mGlu5) is a recognized central nervous system therapeutic target for which several negative allosteric modulator (NAM) drug candidates have or are continuing to be investigated for various disease indications in clinical development. Direct measurement of target receptor occupancy (RO) is extremely useful to help design and interpret efficacy and safety in nonclinical and clinical studies. In the mGlu5 field, this has been successfully achieved by monitoring displacement of radiolabeled ligands, specifically binding to the mGlu5 receptor, in the presence of an mGlu5 NAM using in vivo and ex vivo binding in rodents and positron emission tomography imaging in cynomolgus monkeys and humans. The aim of this study was to measure the RO of the mGlu5 NAM HTL0014242 in rodents and cynomolgus monkeys and to compare its plasma and brain exposure-RO relationships with those of clinically tested mGlu5 NAMs dipraglurant, mavoglurant, and basimglurant. Potential sources of variability that may contribute to these relationships were explored. Distinct plasma exposure-response relationships were found for each mGlu5 NAM, with >100-fold difference in plasma exposure for a given level of RO. However, a unified exposure-response relationship was observed when both unbound brain concentration and mGlu5 affinity were considered. This relationship showed <10-fold overall difference, was fitted with a Hill slope that was not significantly different from 1, and appeared consistent with a simple Emax model. This is the first time this type of comparison has been conducted, demonstrating a unified brain exposure-RO relationship across several species and mGlu5 NAMs with diverse properties. SIGNIFICANCE STATEMENT: Despite the long history of mGlu5 as a therapeutic target and progression of multiple compounds to the clinic, no formal comparison of exposure-receptor occupancy relationships has been conducted. The data from this study indicate for the first time that a consistent, unified relationship can be observed between exposure and mGlu5 receptor occupancy when unbound brain concentration and receptor affinity are taken into account across a range of species for a diverse set of mGlu5 negative allosteric modulators, including a new drug candidate, HTL0014242.

Comparison of models for missing pedigree in single-step genomic prediction.

Pedigree information is often missing for some animals in a breeding program. Unknown-parent groups (UPGs) are assigned to the missing parents to avoid biased genetic evaluations. Although the use of UPGs is well established for the pedigree model, it is unclear how UPGs are integrated into the inverse of the unified relationship matrix (H-inverse) required for single-step genomic best linear unbiased prediction. A generalization of the UPG model is the metafounder (MF) model. The objectives of this study were to derive 3 H-inverses and to compare genetic trends among models with UPG and MF H-inverses using a simulated purebred population. All inverses were derived using the joint density function of the random breeding values and genetic groups. The breeding values of genotyped animals (u2) were assumed to be adjusted for UPG effects (g) using matrix Q2 as u2∗=u2+Q2g before incorporating genomic information. The Quaas-Pollak-transformed (QP) H-inverse was derived using a joint density function of u2∗ and g updated with genomic information and assuming nonzero cov(u2∗,g'). The modified QP (altered) H-inverse also assumes that the genomic information updates u2∗ and g, but cov(u2∗,g')=0. The UPG-encapsulated (EUPG) H-inverse assumed genomic information updates the distribution of u2∗. The EUPG H-inverse had the same structure as the MF H-inverse. Fifty percent of the genotyped females in the simulation had a missing dam, and missing parents were replaced with UPGs by generation. The simulation study indicated that u2∗ and g in models using the QP and altered H-inverses may be inseparable leading to potential biases in genetic trends. Models using the EUPG and MF H-inverses showed no genetic trend biases. These 2 H-inverses yielded the same genomic EBV (GEBV). The predictive ability and inflation of GEBVs from young genotyped animals were nearly identical among models using the QP, altered, EUPG, and MF H-inverses. Although the choice of H-inverse in real applications with enough data may not result in biased genetic trends, the EUPG and MF H-inverses are to be preferred because of theoretical justification and possibility to reduce biases.

Elastic Damping Mechanism Optimization by Indefinite Lagrange Multipliers

The current paper has presented the study on theoretical dependences of the optimization parameter (the degree of transmission transparency) on the design factors of elastic-damping mechanism (EDM) in the transmission for tractors of small traction class (14 kN). The purpose of the study was to obtain a unified functional relationship that integrates early study results. The determination of a unified function was based on reaching a compromise by the method of indefinite Lagrange multipliers. With the help of the function, there have been obtained the dependences of the effect of EDM design factors on transmission processes when performing the main agricultural operations. Using a method based on finding a compromise solution (method of indefinite Lagrange multipliers), there have been substantiated design parameters of the elastic-damping mechanism in the tractor transmission and there has been identified an optimization parameter of the Lagrange function. For a better understanding of the obtained dependence, there has been carried out a three-dimensional visual analysis. Each factor used in the Lagrange function was a physical parameter of an elastic-damping mechanism. The resulting equation has given an opportunity to control mechanism parameters in natural operating conditions. The obtained Lagrange value has demonstrated that the mechanism absorbs 32% of fluctuations in the external traction load of a tractor, when working with the main agricultural tools.

Non-Thermal Melting of Tungsten Under Intense Electronic Excitations

Non-thermal effects caused by ultrafast lasers and swift ions in materials are very intriguing, which is of both scientific interest and technological importance. However, the underlying physics of non-thermal effects on ultrafast process remains unclear and the proposed mechanisms have been controversial. Based on the perturbation approximation under tight-binding theory, the non-thermal effects on tungsten (W) are extensively studied. We demonstrate that the non-thermal effects stemmed from the intense electronic excitations induce dramatic decrease in the melting point of W crystal, as well as non-thermal melting inside the W slab. Our analysis shows that the non-thermal forces are essentially responsible for the drop of melting point of the bulk system. Remarkably, the non-thermal effects combined with surface effects on a W film enhance the ordering of the direction of atomic motion near the surface, preventing melting near the surface area, but leading to non-thermal melting in the interior area of the film. Our work also exhibits a unified relationship between the non-thermal melting and the interatomic forces. This relationship is universal in metals and semiconductors irradiated by ultrafast lasers or fast ions, and has been well established long before.

Parameter Scaling of the Aerodynamic Breakup of the Acoustic Levitated Droplets in an Air Jet Flow

The aerodynamic breakup of the droplet has been intensely studied in this paper. We aim to establish a unified relationship between dimensionless kinematic parameters such as displacement, spreading diameter, Weber number, time, and so on. The breakup characteristics of the acoustic levitated ethanol droplet are experimentally investigated when exposed to an air jet flow. The breakup phenomenons were recorded with a high-speed camera. The breakup characteristics were analyzed, and the physical models of the moving and transforming behaviors were established to explain the breakup mechanisms. We found that the displacement of the windward side of the droplet follows free acceleration rule, with the displacement, acceleration, and time in the dimensionless form. The spreading of the diameter during deformation can also be written in a simple equation as a function of Weber number and displacement. We also discussed more details.