Early stopping monitors global validation loss and halts all parameter updates simultaneously, which is computationally costly for large transformers due to the extended time required for validation inference. We propose \textit{GradES}, a novel gradient-based early stopping approach that operates within transformer components (attention projections and Feed-Forward layer matrices). We found that different components converge at varying rates during fine-tuning for both language and vision-language models. \textit{GradES} tracks the magnitude of gradient changes in backpropagation for these matrices during training. When a projection matrix's magnitude of gradient changes fall below a convergence threshold $\tau$, we exclude that projection matrix from further updates individually, eliminating costly validation passes while allowing slow converging matrices to continue learning. \textit{GradES} speeds up training time by 1.57--7.22$\times$ while simultaneously enhancing generalization through early prevention of overfitting, resulting in 1.2\% higher average accuracy in language tasks and 3.88\% on multimodal benchmarks.

Summary sentences produced by abstractive summarization models may be coherent and comprehensive, but they lack control and rely heavily on reference summaries. The BRIO training paradigm assumes a non-deterministic distribution to reduce the model's dependence on reference summaries, and improve model performance during inference. This paper presents a straightforward but effective technique to improve abstractive summaries by fine-tuning pre-trained language models, and training them with the BRIO paradigm. We build a text summarization dataset for Vietnamese, called VieSum. We perform experiments with abstractive summarization models trained with the BRIO paradigm on the CNNDM and the VieSum datasets. The results show that the models, trained on basic hardware, outperform all existing abstractive summarization models, especially for Vietnamese.

In this paper, we develop a framework for solving inverse deformation problems using the FEniCS Project finite element software. We validate our approach with experimental imaging data acquired from a soft silicone beam under gravity. In contrast with inverse iterative algorithms that require multiple solutions of a standard elasticity problem, the proposed method can compute the undeformed configuration by solving only one modified elasticity problem. This modified problem has a complexity comparable to the standard one. The framework is implemented within an open-source pipeline enabling the direct and inverse deformation simulation directly from imaging data. We use the high-level Unified Form Language (UFL) of the FEniCS Project to express the finite element model in variational form and to automatically derive the consistent Jacobian. Consequently, the design of the pipeline is flexible: for example, it allows the modification of the constitutive models by changing a single line of code. We include a complete working example showing the inverse deformation of a beam deformed by gravity as supplementary material.

{Recognizing human interactions is essential for social robots as it enables them to navigate safely and naturally in shared environments. Conventional robotic systems however often focus on obstacle avoidance, neglecting social cues necessary for seamless human-robot interaction. To address this gap, we propose a framework to recognize human group interactions for socially aware navigation. Our method utilizes color and depth frames from a monocular RGB-D camera to estimate 3D human keypoints and positions. Principal component analysis (PCA) is then used to determine dominant interaction directions. The shoelace formula is finally applied to compute interest points and engagement areas. Extensive experiments have been conducted to evaluate the validity of the proposed method. The results show that our method is capable of recognizing group interactions across different scenarios with varying numbers of individuals. It also achieves high-speed performance, processing each frame in approximately 4 ms on a single-board computer used in robotic systems. The method is implemented as a ROS 2 package making it simple to integrate into existing navigation systems. Source code is available at this https URL

This paper investigates the contribution of business model innovations in improvement of food supply chains. Through a systematic literature review, the notable business model innovations in the food industry are identified, surveyed, and evaluated. Findings reveal that the innovations in value proposition, value creation processes, and value delivery processes of business models are the successful strategies proposed in food industry. It is further disclosed that rural female entrepreneurs, social movements, and also urban conditions are the most important driving forces inducing the farmers to reconsider their business models. In addition, the new technologies and environmental factors are the secondary contributors in business model innovation for the food processors. It is concluded that digitalization has disruptively changed the food distributors models. E-commerce models and internet of things are reported as the essential factors imposing the retailers to innovate their business models. Furthermore, the consumption demand and the product quality are two main factors affecting the business models of all the firms operating in the food supply chain regardless of their positions in the chain. The findings of the current study provide an insight into the food industry to design a sustainable business model to bridge the gap between food supply and food demand.

Wind power as a renewable source of energy, has numerous economic, environmental and social benefits. In order to enhance and control renewable wind power, it is vital to utilize models that predict wind speed with high accuracy. Due to neglecting of requirement and significance of data preprocessing and disregarding the inadequacy of using a single predicting model, many traditional models have poor performance in wind speed prediction. In the current study, for predicting wind speed at target stations in the north of Iran, the combination of a multi-layer perceptron model (MLP) with the Whale Optimization Algorithm (WOA) used to build new method (MLP-WOA) with a limited set of data (2004-2014). Then, the MLP-WOA model was utilized at each of the ten target stations, with the nine stations for training and tenth station for testing (namely: Astara, Bandar-E-Anzali, Rasht, Manjil, Jirandeh, Talesh, Kiyashahr, Lahijan, Masuleh, and Deylaman) to increase the accuracy of the subsequent hybrid model. The capability of the hybrid model in wind speed forecasting at each target station was compared with the MLP model without the WOA optimizer. To determine definite results, numerous statistical performances were utilized. For all ten target stations, the MLP-WOA model had precise outcomes than the standalone MLP model. The hybrid model had acceptable performances with lower amounts of the RMSE, SI and RE parameters and higher values of NSE, WI, and KGE parameters. It was concluded that the WOA optimization algorithm can improve the prediction accuracy of MLP model and may be recommended for accurate wind speed prediction.

This paper proposes an effective computational tool for brittle crack propagation problems based on a combination of a higher-order phase-field model and a non-conforming mesh using a NURBS-based isogeometric approach. This combination, as demonstrated in this paper, is of great benefit in reducing the computational cost of using a local refinement mesh and a higher-order phase-field, which needs higher derivatives of basis functions. Compared with other approaches using a local refinement mesh, the Virtual Uncommon-Knot-Inserted Master-Slave (VUKIMS) method presented here is not only simple to implement but can also reduce the variable numbers. VUKIMS is an outstanding choice in order to establish a local refinement mesh, i.e. a non-conforming mesh, in a multi-patch problem. A phase-field model is an efficient approach for various complicated crack patterns, including those with or without an initial crack path, curved cracks, crack coalescence, and crack propagation through holes. The paper demonstrates that cubic NURBS elements are ideal for balancing the computational cost and the accuracy because they can produce accurate solutions by utilising a lower degree of freedom number than an extremely fine mesh of first-order B-spline elements.

In this paper, we study the Fenchel-Rockafellar duality and the Lagrange duality in the general frame work of vector spaces without topological structures. We utilize the geometric approach, inspired from its successful application by B. S. Mordukhovich and his coauthors in variational and convex analysis. After revisiting coderivative calculus rules and providing the subdifferential maximum rule in vector spaces, we establish conjugate calculus rules under qualifying conditions through the algebraic interior of the function's domains. Then we develop sufficient conditions which guarantee the Fenchel-Rockafellar strong duality. Finally, after deriving some necessary and sufficient conditions for optimal solutions to convex minimization problems, under a Slater condition via the algebraic interior, we then obtain a sufficient condition for the Lagrange strong duality.

In this work, we study two potentials, the single-field and the two-field, from the modified ($R+\gamma R^n$) gravity in D=8 dimensions. From those potentials, we calculate four observable quantities in inflation, including scalar-to-tensor ratio, spectral index, running index and scalar amplitude. Then, we compare them to the experimental data to verify the righteousness of the models. Last but not least, de Sitter conjectures are brought up with these two potentials to investigate that it is possible or not the theory lay in the Landscape of quantum gravity.

One-loop contributions for decay process $H^{\pm} \rightarrow W^{\pm}Z$ within the Two-Higgs-Doublet Model are computed in the general $\mathcal{R}_{\xi}$ gauge, and its phenomenological applications at future muon--TeV colliders are studied in this paper. The analytic results are confirmed by several consistency tests, for example, the $\xi$-independence, the renormalization-scale stability and the ultraviolet finiteness of the one-loop amplitude. We first perform an updated parameter scan of the Type-X THDM in the phenomenological studies. The production of charged Higgs boson pairs at future muon--TeV colliders is investigated through the two processes $\mu^+\mu^- \rightarrow H^+H^- \rightarrow W^{\pm}W^{\mp}Zh$ and $\mu^+\mu^- \rightarrow \gamma\gamma \rightarrow H^+H^- \rightarrow W^{\pm}W^{\mp}Zh$. Both signal events and their significances are evaluated with taking into account the corresponding Standard Model backgrounds. We find that the signal significances can exceed $5\sigma$ at several benchmark points in the viable parameter space of the Type-X THDM.

The ribosomal exit tunnel is the primary structure affecting the release of nascent proteins at the ribosome. The ribosomal exit tunnels from different species have elements of conservation and differentiation in structural and physico-chemical properties. In this study, by simulating the elongation and escape processes of nascent proteins at the ribosomal exit tunnels of four different organisms, we show that the escape process has conserved mechanisms across the domains of life. Specifically, it is found that the escape process of proteins follows the diffusion mechanism given by a simple diffusion model and the median escape time positively correlates with the number of hydrophobic residues and the net charge of a protein for all the exit tunnels considered. These properties hold for twelve distinct proteins considered in two slightly different and improved G\=o-like models. It is also found that the differences in physico-chemical properties of the tunnels lead to quantitative differences in the protein escape times. In particular, the relatively strong hydrophobicity of the E. coli's tunnel and the unusually high number of negatively charged amino acids on the tunnel's surface of H. marismortui lead to substantially slower escapes of proteins at these tunnels than at those of S. cerevisisae and H. sapiens.

This study explores wind energy resources in different locations through the Gulf of Oman and also their future variability due climate change impacts. In this regard, EC-EARTH near surface wind outputs obtained from CORDEX-MENA simulations are used for historical and future projection of the energy. The ERA5 wind data are employed to assess suitability of the climate model. Moreover, the ERA5 wave data over the study area are applied to compute sea surface roughness as an important variable for converting near surface wind speeds to those of wind speed at turbine hub-height. Considering the power distribution, bathymetry and distance from the coats, some spots as tentative energy hotspots to provide detailed assessment of directional and temporal variability and also to investigate climate change impact studies. RCP8.5 as a common climatic scenario is used to project and extract future variation of the energy in the selected sites. The results of this study demonstrate that the selected locations have a suitable potential for wind power turbine plan and constructions.

In the present work we used five different versions of the quark-meson coupling (QMC) model to compute astrophysical quantities related to the GW170817 event and to neutron star cooling process. Two of the models are based on the original bag potential structure and three versions consider a harmonic oscillator potential to confine the quarks. The bag-like models also incorporate the pasta phase used to describe the inner crust of neutron stars. We show that the pasta phase always play a minor or negligible role in all studies. Moreover, while no clear correlation between the models that satisfy the GW170817 constraints and the slope of the symmetry energy is found, a clear correlation is observed between the slope and the fact that the cooling is fast or slow, i.e., fast (slow) cooling is related to higher (lower) values of the slope. We did not find one unique model that can describe, at the same time, GW170817 constraints and give a perfect description of the possible cooling processes.

Active matter systems such as eukaryotic cells and bacteria continuously transform chemical energy to motion. Hence living systems exert active stresses on the complex environments in which they reside. One recurring aspect of this complexity is the viscoelasticity of the medium surrounding living systems: bacteria secrete their own viscoelastic extracellular matrix, and cells constantly deform, proliferate, and self-propel within viscoelastic networks of collagen. It is therefore imperative to understand how active matter modifies, and gets modified by, viscoelastic fluids. Here, we present a two-phase model of active nematic matter that dynamically interacts with a passive viscoelastic polymeric phase and perform numerical simulations in two dimensions to illustrate its applicability. Motivated by recent experiments we first study the suppression of cell division by a viscoelastic medium surrounding the cell. We further show that the self-propulsion of a model keratocyte cell is modified by the polymer relaxation of the surrounding viscoelastic fluid in a non-uniform manner and find that increasing polymer viscosity effectively suppresses the cell motility. Lastly, we explore the hampering impact of the viscoelastic medium on the generic hydrodynamic instabilities of active nematics by simulating the dynamics of an active stripe within a polymeric fluid. The model presented here can provide a framework for investigating more complex dynamics such as the interaction of multicellular growing systems with viscoelastic environments.

We will investigate numerically a seesaw model with $A_4$ flavor symmetry to find allowed regions satisfying the current experimental neutrino oscillation data, then use them to predict physical consequences. Namely, the lightest active neutrino mass has order of $\mathcal{O}(10^{-2})$ eV. The effective neutrino mass $|\langle m\rangle|$ associated with neutrinoless double beta decay is in the range of $[0.002 \;\mathrm{eV},0.038\;\mathrm{eV}]$ and $[0.048\;\mathrm{eV},0.058\;\mathrm{eV}]$ corresponding to the normal and the inverted hierarchy schemes. Other relations among relevant physical quantities are shown, so that they can be determined if some of them are confirmed experimentally. The recent data of the baryon asymmetry of the Universe ($\eta_B$) can be explained via leptogenesis caused by the effect of the renormalization group evolution on the Dirac Yukawa couplings, provided the right handed neutrino mass scale $M_0$ is ranging from $\mathcal{O}(10^8)$ GeV to $\mathcal{O}(10^{12})$ GeV for $\tan\beta =3$. This allowed $M_0$ range distinguishes with the scale of $\mathcal{O}(10^{13})$ GeV concerned by other effects that also generate the consistent $\eta_B$ from leptogenesis. The branching ratio of the decay $\mu \rightarrow\,e\gamma$ may reach the future experimental sensitivity in the very light values of $M_0$. Hence, it will be inconsistent with the $M_0$ range predicted from the $\eta_B$ data whenever this decay is detected experimentally.

Sphaleron electroweak phase transition (EWPT) is calculated in two phase transition stages, thereby showing that the twin (or double) bubble nucleation structure of the phase transition and gravitational wave is in the investigation area of future detectors. With $v^2=v^2_1+v^2_2$ ($v_1$ and $v_2$ are two vacuum average values (VEV)), the parameter $\tan\beta=v_2/v_1$, is the ratio between two VEVs although it does not affect the strength of EWPT but affect the sphaleron energy. However, it only causes this energy to increase slightly. As $a=v^2/v_2^2$ increases, the maximum difference of sphaleron energy in one stage is about $6.15$ TeV. $a$ affects the expansion of bubbles during two phase transitions. The more $a$ increases, the more the expansion of two bubbles is at the same time. This ratio does not greatly affect the sphaleron energy but has an impact on gravitational waves. The larger the masses of the charged Higgs particles are, the greater the gravitational wave energy density ($\Omega h^2$) is. When the frequency is in the range $0-1.2$ mHz, $\Omega h^2$ will has a maximum value in the range $10^{-12}-10^{-11}$ for all values of $a$ so this can be detected in the future.

This paper describes the canonical quantization of the U(1) gauge field across all four regions in the Rindler coordinates in the Lorentz-covariant gauge. Concretely, in the four regions (future, past, left and right Rindler-wedges) in the Rindler coordinates, the gauge-fixed Lagrangian in the Lorentz-covariant gauge is obtained, which is composed of the U(1) gauge field, the B-field and ghost fields. Since the U(1) gauge and B-fields are decoupled from the ghost fields by the property of the U(1) gauge theory, the U(1) gauge field and the B-field are examined in this study. Then, by solving the equations of motion obtained from the gauge-fixed Lagrangian, the solutions of each mode of the U(1) gauge field and the B-field can be obtained. Following this, with the Klein-Gordon inner-product defined in the Rindler coordinates, the normalization constants of each of those mode-solutions are determined. Subsequently, formulating the canonical commutation relations of the U(1) gauge field and its canonical conjugate momentum, the equal-time commutation relations of the coefficient of each mode-solution in each direction of the U(1) gauge field in each region of the Rindler coordinates are obtained. From these, it can be seen that those coefficients have physical meaning as creation/annihilation operators. The polarization vectors in each region of the Rindler coordinates are also given in this study.

We establish an improved version of the Moser-Trudinger inequality in the hyperbolic space $\mathbb H^n$, $n\geq 2$. Namely, we prove the following result: for any 0 \leq \lambda &lt; \left(\frac{n-1}n\right)^n, then we have $$ \sup_{\substack{u\in C_0^\infty(\mathbb H^n) \int_{\mathbb H^n} |\nabla_g u|_g^n d\text{Vol}_g -\lambda \int_{\mathbb H^n} |u|^n d\text{ Vol}_g \leq 1}} \int_{\mathbb H^n} \Phi_n(\alpha_n |u|^{\frac{n}{n-1}}) d\text{ Vol}_g < \infty, $$ where $\alpha_n = n \omega_{n-1}^{\frac1{n-1}}$, $\omega_{n-1}$ denotes the surface area of the unit sphere in $\mathbb R^n$ and $\Phi_n(t) = e^t -\sum_{j=0}^{n-2}\frac{t^j}{j!}$. This improves the Moser-Trudinger inequality in hyperbolic spaces obtained recently by Mancini and Sandeep, by Mancini, Sandeep and Tintarev and by Adimurthi and Tintarev. In the limiting case $\lambda =(\frac{n-1}n)^n$, we prove a Moser-Trudinger inequality with exact growth in $\mathbb H^n$, $$ \sup_{\substack{u\in C_0^\infty(\mathbb H^n) \int_{\mathbb H^n} |\nabla_g u|_g^n d\text{ Vol}_g -(\frac{n-1}n)^n \int_{\mathbb H^n} |u|^n d\text{ Vol}_g \leq 1}} \frac{1}{\int_{\mathbb H^n} |u|^n d\text{ Vol}_g}\int_{\mathbb H^n} \frac{\Phi_n(\alpha_n |u|^{\frac{n}{n-1}})}{(1+ |u|)^{\frac n{n-1}}} d\text{ Vol}_g < \infty. $$ This improves the Moser-Trudinger inequality with exact growth in $\mathbb H^n$ established by Lu and Tang.

The Jost function method is extended within the framework of the RPA theory to find poles on the complex energy plane that exhibit complex RPA eigenenergies. Poles corresponding to the RPA excited states such as the giant resonance of $^{16}$O electric dipole excitations were successfully found on the complex energy plane. Although the giant resonance has been known as a single resonance with large strength and width, it is found that, at least within the RPA framework, the $^{16}$O electric dipole giant resonance is formed by multiple poles, each of which is an independent pole with different widths, origins, response properties to residual interactions, and components structures of the density fluctuation.