Search

Recently, learned image compression methods based on entropy minimization have achieved superior results compared with conventional image codecs such as BPG and JPEG2000. However, they leverage single Gaussian models, which have a limited ability to approximate various irregular distributions of trans-formed latent representations, resulting in suboptimal coding efficiency. Fur-thermore, existing methods focus on constructing effective entropy models, rather than utilizing modern architectural techniques. In this paper, we pro-pose a novel joint learning scheme called JointIQ-Net that incorporates image compression and quality enhancement technologies with improved entropy minimization based on a newly adopted Gaussian mixture model. We also exploit global context to estimate the distributions of latent representations precisely. The results of extensive experiments demonstrate that JointIQ-Net achieves remarkable performance improvements in terms of coding efficiency compared with existing learned image compression methods and conventional codecs. To the best of our knowledge, ours is the first learned image compres-sion method that outperforms VVC intra-coding in terms of both PSNR and MS-SSIM.

Abstract Background: The collection of substantial amounts of electroencephalogram (EEG) data is typically time-consuming and labor-intensive, which adversely impacts the development of decoding models with strong generalizability, particularly when the available data is limited. Utilizing sufficient EEG data from other subjects to aid in modeling the target subject presents a potential solution, commonly referred to as domain adaptation. Most current domain adaptation techniques for EEG decoding primarily focus on learning shared feature representations through domain alignment strategies. Since the domain shift cannot be completely removed, target EEG samples located near the edge of clusters are also susceptible to misclassification. Methods: We propose a novel adaptive deep feature representation (ADFR) framework to improve the cross-subject EEG classification performance through learning transferable EEG feature representations. Specifically, we first minimize the distribution discrepancy between the source and target domains by employing maximum mean discrepancy (MMD) regularization, which aids in learning the shared feature representations. We then utilize the instance-based discriminative feature learning (IDFL) regularization to make the learned feature representations more discriminative. Finally, the entropy minimization (EM) regularization is further integrated to adjust the classifier to pass through the low-density region between clusters. The synergistic learning between above regularizations during the training process enhances EEG decoding performance across subjects. Results: The effectiveness of the ADFR framework was evaluated on two public motor imagery (MI)-based EEG datasets: BCI Competition III dataset 4a and BCI Competition IV dataset 2a. In terms of average accuracy, ADFR achieved improvements of 3.0% and 2.1%, respectively, over the state-of-the-art methods on these datasets. Conclusions: The promising results highlight the effectiveness of the ADFR algorithm for EEG decoding and show its potential for practical applications.

Background: The collection of substantial amounts of electroencephalogram (EEG) data is typically time-consuming and labor-intensive, which adversely impacts the development of decoding models with strong generalizability, particularly when the available data is limited. Utilizing sufficient EEG data from other subjects to aid in modeling the target subject presents a potential solution, commonly referred to as domain adaptation. Most current domain adaptation techniques for EEG decoding primarily focus on learning shared feature representations through domain alignment strategies. Since the domain shift cannot be completely removed, target EEG samples located near the edge of clusters are also susceptible to misclassification. Methods: We propose a novel adaptive deep feature representation (ADFR) framework to improve the cross-subject EEG classification performance through learning transferable EEG feature representations. Specifically, we first minimize the distribution discrepancy between the source and target domains by employing maximum mean discrepancy (MMD) regularization, which aids in learning the shared feature representations. We then utilize the instance-based discriminative feature learning (IDFL) regularization to make the learned feature representations more discriminative. Finally, the entropy minimization (EM) regularization is further integrated to adjust the classifier to pass through the low-density region between clusters. The synergistic learning between above regularizations during the training process enhances EEG decoding performance across subjects. Results: The effectiveness of the ADFR framework was evaluated on two public motor imagery (MI)-based EEG datasets: BCI Competition III dataset 4a and BCI Competition IV dataset 2a. In terms of average accuracy, ADFR achieved improvements of 3.0% and 2.1%, respectively, over the state-of-the-art methods on these datasets. Conclusions: The promising results highlight the effectiveness of the ADFR algorithm for EEG decoding and show its potential for practical applications. [ABSTRACT FROM AUTHOR]

Graph clustering is one of the most fundamental tasks in graph learning. Recently, numerous graph clustering models based on dual network (Auto-encoder+Graph Neural Network(GNN)) architectures have emerged and achieved promising results. However, we observe several limitations in the literature: 1) simple graph neural networks that fail to capture the intricate relationships between nodes are used for graph clustering tasks; 2) heterogeneous information is inadequately interacted and merged; and 3) the clustering boundaries are fuzzy in the feature space. To address the aforementioned issues, we propose a novel graph clustering model named Dynamic Graph Attention-guided Graph Clustering with Entropy Minimization self-supervision(DGAGC-EM). Specifically, we introduce DGATE, a graph auto-encoder based on dynamic graph attention, to capture the intricate relationships among graph nodes. Additionally, we perform feature enhancement from both global and local perspectives via the proposed Global-Local Feature Enhancement (GLFE) module. Finally, we propose a self-supervised strategy based on entropy minimization theory to guide network training process to achieve better performance and produce sharper clustering boundaries. Extensive experimental results obtained on four datasets demonstrate that our method is highly competitive with the SOTA methods. The figure presents the overall framework of proposed Dynamic Graph Attention-guided Graph Clustering with Entropy Minimization selfsupervision(DGAGC-EM). Specifically, the Dynamic Graph Attetion Auto-Encoder Module is our proposed graph auto-encoder based on dynamic graph attention, to capture the intricate relationships among graph nodes. The Auto-Encoder Module is a basic autoencoder with simple MLPs to extract embeddings from node attributes. Additionally, the proposed Global-Local Feature Enhancement (GLFE) module perform feature enhancement from both global and local perspectives. Finally, the proposed Self-supervised Module guide network training process to achieve better performance and produce sharper clustering boundaries [ABSTRACT FROM AUTHOR]

Entropy generation is an imperative feature of every thermal transportation processes. It helps to decrease irreversibility factor in a system. Therefore, present study investigates entropy generation analysis for a peristaltic motion of fourth-grade fluid in a symmetric channel in the presence of an induced magnetic field. Effects of thermal radiation, viscous dissipation, and thermal absorption/generation are considered. Slip condition is also imposed at the upper wall of the channel. Furthermore, the liquid is pondered to possess variable viscosity which displays exponential alteration over the width of the channel. Lubrication approximation theory is used to simplify partial differential equations that are strongly nonlinear. The perturbation technique has been used to yield the solution for momentum and stream function. However, the energy equation is solved numerically. The impacts of numerous embedded dimensionless substantial parameters on liquid flow are exhibited through graphs. Outcomes divulge that entropy generation enhances with an escalation of thermal conductivity parameter, however it diminutions with an enhancement of magnetic parameter. Such outcomes benefits in biomedical sciences. [ABSTRACT FROM AUTHOR]

Recently, learned image compression methods based on entropy minimization have achieved superior results compared with conventional image codecs such as BPG and JPEG2000. However, they leverage single Gaussian models, which have a limited ability to approximate various irregular distributions of transformed latent representations, resulting in suboptimal coding efficiency. Furthermore, existing methods focus on constructing effective entropy models, rather than utilizing modern architectural techniques. In this paper, we propose a novel joint learning scheme called JointIQ‐Net that incorporates image compression and quality enhancement technologies with improved entropy minimization based on a newly adopted Gaussian mixture model. We also exploit global context to estimate the distributions of latent representations precisely. The results of extensive experiments demonstrate that JointIQ‐Net achieves remarkable performance improvements in terms of coding efficiency compared with existing learned image compression methods and conventional codecs. To the best of our knowledge, ours is the first learned image compression method that outperforms VVC intra‐coding in terms of both PSNR and MS‐SSIM. [ABSTRACT FROM AUTHOR]

When there are no constraints upon the solutions of the equation 푨 ⁢ 흃 = 풚 {{\boldsymbol{A}}{\boldsymbol{\xi}}={\boldsymbol{y}}} , where 푨 {{\boldsymbol{A}}} is a K × N - {K\times N-} matrix, 흃 ∈ ℝ N {{\boldsymbol{\xi}}\in{\mathbb{R}}^{N}} and 풚 ∈ ℝ K {{\boldsymbol{y}}\in{\mathbb{R}}^{K}} a given vector, the description of the set of solutions as 풚 {{\boldsymbol{y}}} varies in ℝ K {{\mathbb{R}}^{K}} is well known. But this is not so when the solutions are required to satisfy 흃 ∈ 풦 ⁢ ∏ i ≤ j ≤ N [ a j , b j ] {{\boldsymbol{\xi}}\in{\mathcal{K}}\prod_{i\leq j\leq N}[a_{j},b_{j}]} , for finite a j ≤ b j : 1 ≤ j ≤ N {a_{j}\leq b_{j}:1\leq j\leq N} . To solve this problem we bring in a strictly convex, Fermi-Dirac entropy function Ψ ⁢ ( 흃 ) {\Psi({\boldsymbol{\xi}})} , and find the solution as a ⁢ r ⁢ g ⁢ m ⁢ i ⁢ n ⁢ { Ψ ⁢ ( 흃 ) : 흃 ∈ 풦 , 푨 ⁢ 흃 = y } {argmin\{\Psi({\boldsymbol{\xi}}):{\boldsymbol{\xi}}\in{\mathcal{K}},\,{% \boldsymbol{A}}{\boldsymbol{\xi}}=y\}} . If λ denotes the Lagrange multipliers of the optimization problem, we study the properties of the parametric surface 흀 → 흃 ⁢ ( 흀 ) {{\boldsymbol{\lambda}}\to{\boldsymbol{\xi}}({\boldsymbol{\lambda}})} in the geometry on 풦 {{\mathcal{K}}} defined by the Hessian metric derived from Ψ ⁢ ( 흃 ) {\Psi({\boldsymbol{\xi}})} . In particular, we prove that the surface 흀 → 흃 ⁢ ( 흀 ) {{\boldsymbol{\lambda}}\to{\boldsymbol{\xi}}({\boldsymbol{\lambda}})} is contained in ker ( 푨 ) ⊥ {\ker({\boldsymbol{A}})^{\perp}} in the Hessian metric derived from Ψ. [ABSTRACT FROM AUTHOR]

The imaging of aerial targets using Inverse Synthetic Aperture Radar (ISAR) is affected by micro-Doppler effects resulting from localized micromotions, such as rotation and vibration. These effects introduce additional Doppler frequency modulation into the echo, leading to spectral broadening. Under ultrahigh-resolution conditions, these micromotions interfere with the focusing process of subject scatterers, resulting in images with poor focus showing significantly reduced quality. Furthermore, micro-Doppler signals exhibit temporal variability and nonstationary characteristics, posing difficulties in their estimation and differentiation from the echo. To address these challenges, this paper proposes a nonparametric method based on Variational Mode Decomposition (VMD) and mode optimization to separate the echo of the subject from micro-Doppler components. This separation is achieved by utilizing differences in their respective time-frequency distributions. This methodology mitigates the effect of micro-Doppler signals on the echo and obtains imaging results of a drone with ultrahigh-resolution. The VMD algorithm is introduced and subsequently extended to the complex domain. The method entails the decomposition of the ISAR echo along the azimuth direction into several mode functions distributed uniformly across the Doppler sampling bandwidth. Subsequently, image entropy indices are employed to optimize the decomposition parameters and select the imaging modes. This ensures the effective suppression of micro-Doppler signals and preservation of the subject echo. Compared to existing methods based on Empirical Mode Decomposition (EMD) and Local Mean Decomposition (LMD), the proposed method exhibits superior performance in suppressing image blurring caused by micro-Doppler effects while ensuring complete retention of fuselage details. Furthermore, the effectiveness and advantages of the proposed method are validated through simulations and processing of ultrawideband microwave photonic data obtained from drone measurements.

Space targets move in orbit at a very high speed, so in order to obtain high-quality imaging, high-speed motion compensation (HSMC) and translational motion compensation (TMC) are required. HSMC and TMC are usually adjacent, and the residual error of HSMC will reduce the accuracy of TMC. At the same time, under the condition of low signal-to-noise ratio (SNR), the accuracy of HSMC and TMC will also decrease, which brings challenges to high-quality ISAR imaging. Therefore, this paper proposes a joint ISAR motion compensation algorithm based on entropy minimization under low-SNR conditions. Firstly, the motion of the space target is analyzed, and the echo signal model is obtained. Then, the motion of the space target is modeled as a high-order polynomial, and a parameterized joint compensation model of high-speed motion and translational motion is established. Finally, taking the image entropy after joint motion compensation as the objective function, the red-tailed hawk–Nelder–Mead (RTH-NM) algorithm is used to estimate the target motion parameters, and the joint compensation is carried out. The experimental results of simulation data and real data verify the effectiveness and robustness of the proposed algorithm. [ABSTRACT FROM AUTHOR]

Copyright of Environmental & Social Management Journal / Revista de Gestão Social e Ambiental is the property of Environmental & Social Management Journal and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

The effects of different nanoparticle shapes and uncertainty in the nanofluid thermophysical properties on the optimal operating conditions of MHD flow of Al2O3/water nanofluid through a horizontal microchannel with a porous medium considering hydrodynamic slip, suction/injection and thermal radiation were investigated. The momentum and heat transfer equations were solved analytically using the methods of undetermined coefficients and variation of constants, respectively. From the exact solutions of the velocity and temperature fields, the global entropy S and Nusselt number Nu were computed. The impacts of hydrodynamic slip α , Biot number Bi , nanoparticle concentration ϕ and Darcy number Da on entropy production and heat transport were investigated. The results revealed that optimum values of Bi and α with minimum global entropy and maximum heat transport were achieved for symmetric slip conditions and asymmetric heat transfer. The platelet shape of nanoparticles was the most effective to achieve the optimum conditions with the lowest minimum value of S , while the blade shape was the most effective to reach the optimum conditions with the highest maximum value of heat transport. Thus, optimum values of both Biot number of bottom plate equal to 0.01 and slip equal to 0.045 with the smallest values of S were achieved for the platelet shape. Also, optimum slip value of 0.15 with the largest maximum Nu at top plate of 5.13 was achieved for the blade shape. On the other hand, when ϕ increased from 0 to 0.045, S always decreased and Nu always increased. The greatest decrease of entropy from 0.133 to 0.088 (33%) occurred for the platelet shape, while the greatest increase of Nu at top plate from 4.96 to 5.57 (12.3%) occurred for the blade shape. When ϕ was varied from 0 to 0.01, S decreased 9.2% for the platelet shape compared to spherical shapes, and Nu at top plate increased 2.6% for the blade shape compared to spherical shapes. The results also indicated that the greatest variations of optimum operating conditions occurred when the experimental correlations of viscosity and thermal conductivity were used compared to theoretical correlations. This is because the estimated values of viscosity and conductivity using the different theoretical correlations differ very little from each other. Thus, the maximum value of Nu at top plate increased from 5.067 for SM1 model to 5.092 for SM6 model (0.5%), while it increased from 4.96 for EM3 model to 5.17 for EM6 model (4.2%). Finally, the effects of Al2O3, Cu and TiO2 nanoparticles in water as base fluid on the optimum conditions were investigated. Both the lowest entropy production and the highest heat transfer were reached for Cu nanoparticles. When α was varied, the minimum value of S achieved for Cu was 0.47 and 0.64% lower than the minimum value of TiO2 and Al2O3, respectively. Also, the maximum value of Nu achieved for Cu improved by approximately 0.2 and 0.4% compared to Al2O3 and TiO2, respectively. [ABSTRACT FROM AUTHOR]

Metallographic image segmentation is a core task towards the automation of metallographic analysis. Currently, the most advanced methods for this task generally employ supervised deep learning segmentation models that require a great number of pixel-level annotated images, while the annotation process is time-consuming and labor-intensive. In order to address this issue, a semi-supervised model called Con2Net is proposed in this work, which leverages unlabeled data to improve performance for metallographic image segmentation. The Con2Net model adopts a multi-decoder architecture, which enforces consistency constraint between each decoder’s output and other decoders’ soft pseudo labels produced by output sharpening. In addition, to mitigate the negative impact caused by sharpening on false predicted pixels, we adopt the sharpening operation only to accurately predicted pixels. For labeled pixels, comparing ground truth to filter out correctly predicted pixels is the simplest and most effective approach. For unlabeled pixels, a contrastive learning module is introduced, which encourages the model to have better intra-class compactness and inter-class dispersion in the feature space. Based on that, pseudo-labels for unlabeled pixels are obtained by calculating the maximum similarity of feature vectors, and then the accurately predicted unlabeled pixels could be filtered out. We conduct experiments on two public datasets for metallographic image segmentation, comparing the proposed Con2Net model with five state-of-the-art semi-supervised segmentation models on three semi-supervised data partition protocols. The results demonstrate that Con2Net not only outperforms the supervised baseline by a significant margin, but also achieves superior segmentation performance compared to other five semi-supervised models. Our source code is available at https://github.com/Siiimon2423/Con2Net.

In domain adaptation, entropy minimization is widely used. However, entropy minimization will bring negative transfer when the pseudo-labels are inconsistent with the real labels. We hope to increase pseudo-label accuracy to counter negative transfer in entropy minimization. To this end, we introduce domain adversarial training into entropy minimization. Furthermore, we consider the misalignment caused by domain adversarial training under severe label shift. Therefore, we propose method called entropy minimization and domain adversarial training guided by label distribution similarity (EMALDS). Through domain adversarial training which focus more on class-aligned divergence, our method improves pseudo-label accuracy and reduce negative transfer in entropy minimization. Extensive experiments demonstrate the effectiveness and robustness of our proposed method. [ABSTRACT FROM AUTHOR]

This research looks at the incompressible, viscous, steady-state laminar stagnation-point-flow of a Carreau ternary-hybrid nano-liquid towards a convectively heated expanding surface, taking into account first-order slippage and Ohmic dissipation. Viscous dissipation and the effects of Lorentz forces are also considered. It is discussed how thermal radiation and heat absorption/generation contribute to the heat transport process. Minimizing entropy during the transport of a Carreau ternary hybrid nano-liquid has also been studied. In addition, convective circumstances are used conceptually in the numerical solution of the current model. tiny-particles of silver (Ag), molybdenum disulfide (MoS2), and multi-wall carbon nanotubes (MWCNT) are analyzed in this work’s flow study. As a base fluid, carboxymethyl cellulose (CMC-water) is used. With the use of appropriate similarity trans-formations, the standard model equations are transformed into dimensionless form. Finding solutions for momentum and heat fields using the Runge–Kutta-Fehlberg technique and a shooting strategy. The figures depict a wide range of characteristics, including fluid flow velocity, temperature, skin friction, the Nusselt number, entropy minimization, and the Bejan number. Key results from the present model include the fact that the velocity curve flattens down as Weissenberg number increases.

Abstract In this study, we consider a steady-state viscous,laminar Carreau ternary-hybrid nanofluid flow over a stretching sheet with first-order slip and Ohmic dissipation.Further, we consider the effects of suction/injection, oblique Lorentz force, and a temperature jump boundary condition. We investigate here how to transport a Carreau ternary hybrid nano-fluid while minimizing the entropy it generates. We analyze in this article the flow of molybdenum disulfide (MoS2), silver (Ag), and multi-wall carbon nanotubes (MWCNT). Carboxymethyl cellulose (CMC-water) is the observed fluid of choice. Several industries could benefit from this research, including those dealing with the equations of the model are converted to a dimensionless form by appropriate similarity transformations. Using a shooting strategy with Runge-Kutta-Fehlberg and Secant methods the velocity and the temperature fields are obtained. The physical quantities such as the fluid flow rate, the temperature, the skin friction, the Nusselt number, the entropy minimization, and the Bejan number are presented graphically and analyzed. In our Carreau ternary hybrid nanofluid model it is found that an increase in the Weissenberg number is to decrease the velocity field. We also noticed a significant effect of the Brinkmann number which has a deleterious effect on the entropy-minimization and Bejan-number profiles. [ABSTRACT FROM AUTHOR]

This article proposes the entropy of m 2 (m is the stochastic discount factor) as a metric to evaluate asset-pricing models. We develop a bound on the entropy of m 2 when m correctly prices a finite number of returns and consider models that pass the lower bound on m , yet fail the lower bound on m 2. Interpreting our results, we elaborate on the distinction between the entropy of m 2 versus the entropy of m. We further show that the entropy of m 2 represents an upper bound on the expected excess (log) return of the security with the payoff of m. [ABSTRACT FROM AUTHOR]

Deep learning models have achieved remarkable success in medical imaging analysis. However, existing methods are primarily focused on supervised learning, which requires a massive amount of training data. Recent studies have explored semi-supervised learning approaches to address this issue, where data augmentation was applied to unlabeled data. However, there are still two unsolved challenges in applying data augmentation to unlabeled medical images: it can i) result in the lesion features loss and ii) reduce the discriminability of prediction results. Thus, in this work, weak data augmentation is applied to unlabeled data to avoid losing lesions features. Also, we propose nuclear-norm maximization to achieve entropy minimization without losing prediction diversity. Experimental results on two public datasets show that the proposed method outperforms the compared models. [ABSTRACT FROM AUTHOR]

There is a growing interest in exploring interactions at root–soil interface in natural and agricultural ecosystems, but an entropy-based understanding of these dynamic rhizosphere processes is lacking. We have developed a new conceptual model of rhizosphere regulation by localized nutrient supply using thermodynamic entropy. Increased nutrient-use efficiency is achieved by rhizosphere management based on self-organization and minimized entropy via equilibrium attractors comprising (i) optimized root strategies for nutrient acquisition and (ii) improved information exchange related to root–soil–microbe interactions. The cascading effects through different hierarchical levels amplify the underlying processes in plant–soil system. We propose a strategy for manipulating rhizosphere dynamics and improving nutrient-use efficiency by localized nutrient supply with minimization of entropy to underpin sustainable food/feed/fiber production. Despite the inherently complex dynamics of the rhizosphere, it can be engineered. There are still many knowledge gaps regarding the rhizosphere and its engineering and how it can enhance the agricultural production. The models based on entropy can promote the understanding and the engineering of various components of the rhizosphere, up scaling it to the plant–soil system and other higher level systems. Engineering the rhizosphere by emphasizing and harnessing the heterogeneous (localized) nutrient supply to decrease entropy is a promising approach to achieve green, sustainable agriculture by growing more produce with less resources. [ABSTRACT FROM AUTHOR]

The amalgamation of microorganisms in the nanofluid is significant in beautifying the thermal conductivity of several systems, such as microfluid devices, chip-shaped microdevices, and enzyme biosensors. The current investigation studies mixed convective flow of the entropy minimization of unsteady MHD tangent hyperbolic nanoliquid because a stretching surface has motile density via convective and slip conditions. For the novelty of this work, the variable transport characteristics caused by dynamic viscosity, thermal conductivity, nanoparticle mass permeability, and microbial organism diffusivity are considered. It is considered that the vertical sheet studying the flow. By using the appropriate alteration, the governing equations for the most recent flow analysis were altered into a non-dimension relation. Through MATLAB Software bvp4c, the PDE model equations have been made for these transformed equations. Engineering-relevant quantities against various physical variables include force friction, Nusselt number, Sherwood number, and microorganism profiles. The results showed good consistency compared to the current literature. Moreover, these outcomes revealed that augmentation in the magnitude of the magnetic field and velocity slip parameter declines the velocity profile. The reverse impact is studied in We. In addition, heat transfer is typically improved by the influence of thermal radiation parameters, Brownian movement, and thermophoretic force. The physical interpretation has existed through graphical and tabular explanations.

The present study investigates the Marangoni convective fluid flow over a rotating disk with an inclined magnetic field and in the presence of a nonuniform heat source when the disk moves upward/downward with nonconstant velocity with the incorporation of the second law of thermodynamics. The Keller‐box method is applied to the reduced system of equations to draw graphical illustrations. The study of these illustrations to examine the effects of involved pertinent parameters, like, magnetic field, Marangoni number, angle of inclination, vertical disk movement parameter, heat source, and disk rotation, on velocity and temperature profiles, reveals some interesting findings. From the analysis, it can be concluded that the skin friction coefficient increases with more angle of inclination and the Marangoni number with the reverse trend in case of vertical disk movement. Also, the Marangoni number and vertical disk motion diminish the Nusselt number with a positive effect in the case of more angle of inclination. The rate of entropy generation is enhanced with the temperature ratio parameter while it diminishes with the inclined magnetic field of any strength. The current study in its reduced form is in excellent agreement with earlier published work to ensure the validity of the used numerical scheme. [ABSTRACT FROM AUTHOR]

In real industrial scenarios, deep learning-based fault diagnosis has been a popular topic lately. Unfortunately, the source-trained model typically usually underperforms in target domain owning to changeable working conditions. To resolve this problem, a novel self-supervised bi-classifier adversarial transfer learning (SBATL) network by introducing self-supervised learning (SSL) and class-conditional entropy minimization is presented. Concretely, the SBATL is made up of a feature extractor, a discrepancy detector of two classifiers, and a clustering metric based on SSL, which jointly conducts self-supervised and supervised optimization in a two-stream training procedure. In the self-supervised stream, target pseudo labels obtained by SSL are used to construct the topological clustering metric for target feature optimization. In the supervised stream, the feature extractor and classifiers compete with each other in adversarial training, which bridges the discrepancy between two classifiers. Additionally, the class-conditional entropy minimization of target domain is further embedded into both streams to amend the decision boundaries of two classifiers to pass low-density regions. The results indicate that the SBATL gets better cross-domain fault diagnosis performances when compared with other popular methods. • A novel self-supervised bi-classifier adversarial transfer learning is proposed. • A new two-stream strategy is developed to achieve class-level domain adaptation. • With hyperparameter grid search and class-conditional entropy minimization, the model converges stably. • Extensive experiments are constructed to prove the excellent generalization of the proposed method. [ABSTRACT FROM AUTHOR]

Domain adaptation is an important topic due to its capability in transferring knowledge from source domain to target domain. However, many existing domain adaptation methods primarily concentrate on aligning the data distributions between the source and target domains, often neglecting discriminative feature learning. As a result, target samples with low confidence are embedded near the decision boundary, where they are susceptible to being misclassified, resulting in negative transfer. To address this problem, a novel A daptive C enter- A ligned N etwork dubbed ACAN is proposed for unsupervised domain adaptation in this work. The main innovations of ACAN are fourfold. Firstly, it is a plug-and-play module and can be easily incorporated into any domain alignment methods without increasing the model complexity and computational burden. Secondly, in contrast to conventional softmax plus cross-entropy loss, angular margin loss is called to enhance the discrimination power for classifier. Thirdly, entropy regularization is exploited to highlight the probability of potential related class, which renders our learned feature representation far away from the decision boundary. Fourthly, to improve the discriminative capacity of model to the target domain, we propose to align the target domain samples to the corresponding class center via pseudo labels. Incorporating ACAN, the performance of baseline domain alignment methods is significantly improved. Extensive ablation and comparison experiments on four widely adopted databases demonstrate the effectiveness of our ACAN. Code is available at: https://github.com/Cloudfly-Z/ACAN [ABSTRACT FROM AUTHOR]

Clustering is one of the fundamental tasks in computer vision and pattern recognition. Recently, deep clustering methods (algorithms based on deep learning) have attracted wide attention with their impressive performance. Most of these algorithms combine deep unsupervised representation learning and standard clustering together. However, the separation of representation learning and clustering will lead to suboptimal solutions because the two-stage strategy prevents representation learning from adapting to subsequent tasks (e.g., clustering according to specific cues). To overcome this issue, efforts have been made in the dynamic adaption of representation and cluster assignment, whereas current state-of-the-art methods suffer from heuristically constructed objectives with the representation and cluster assignment alternatively optimized. To further standardize the clustering problem, we audaciously formulate the objective of clustering as finding a precise feature as the cue for cluster assignment. Based on this, we propose a general-purpose deep clustering framework, which radically integrates representation learning and clustering into a single pipeline for the first time. The proposed framework exploits the powerful ability of recently developed generative models for learning intrinsic features, and imposes an entropy minimization on the distribution of the cluster assignment by a dedicated variational algorithm. The experimental results show that the performance of the proposed method is superior, or at least comparable to, the state-of-the-art methods on the handwritten digit recognition, fashion recognition, face recognition, and object recognition benchmark datasets. [ABSTRACT FROM AUTHOR]

Geotagging images of interest are increasingly important to law enforcement, national security, and journalism. Today, many images do not carry location tags that are trustworthy and resilient to tampering; and landmark-based visual clues may not be readily present in every image, especially in those taken indoors. In this paper, we exploit an environmental signature from the power grid, the electric network frequency (ENF) signal, which can be inherently captured in a sensing stream at the time of recording and carries useful time–location information. Compared to the recent art of extracting ENF traces from audio and video recordings, it is very challenging to extract an ENF trace from a single image. We address this challenge by first mathematically examining the impact of the ENF embedding steps such as electricity to light conversion, scene geometry dilution of radiation, and image sensing. We then incorporate the verified parametric models of the physical embedding process into our proposed entropy minimization method. The optimized results of the entropy minimization are used for creating a two-level ENF presence–classification test for region-of-capturing localization. It identifies whether a single image has an ENF trace; if yes, whether it is at 50 or 60 Hz. We quantitatively study the relationship between the ENF strength and its detectability from a single image. This paper is the first comprehensive work to bring out a unique forensic capability of environmental traces that shed light on an image’s capturing location. [ABSTRACT FROM AUTHOR]

Abstract Range alignment is an essential procedure in the translation motion compensation of inverse synthetic aperture radar imaging. Global optimization or maximum-correlation-based algorithms have been used to realize range alignment. However, it is still challenging to achieve range alignment in low signal-to-noise ratio scenarios, which are common in inverse synthetic aperture radar imaging. In this paper, a novel anti-noise range alignment approach is proposed. In this new method, the target motion is modeled as a uniformly accelerated motion during a short sub-aperture time. Minimum entropy optimization is implemented to estimate the motion parameters in each sub-aperture. These estimated parameters can be used to align the profiles of the current sub-aperture. Once the range profiles of each sub-aperture are aligned, the non-coherent accumulation gain is obtained by averaging all profiles in each sub-aperture, which can be used as valuable information. The accumulation and correlation method is applied to align the average range profiles of each sub-aperture because the former step focuses mainly on alignment within the sub-apertures. Experimental results based on simulated and real measured data demonstrate the effectiveness of the proposed algorithm in low signal-to-noise ratio scenarios.

Automated segmentation of the optic disc (OD) and optic cup (OC) from different datasets plays an important role in the diagnosis of glaucoma and greatly saves human resources in both data annotation and image segmentation. However, the domain shift between different datasets suppresses the generalization ability of the segmentation network, especially damaging the performance of segmentation in the target domain, which is unlabeled. Therefore, using a transfer learning algorithm or domain adaptation method to enhance the migration ability of segmentation models has become an essential step and has attracted the attention of many researchers. In this paper, we propose an unsupervised domain adaptation network, called the Minimizing-entropy and Fourier Domain Adaptation network (MeFDA), to narrow the discrepancy between the source and target domains and prevent the degradation of segmentation performance. First, we perform adversarial optimization on the entropy maps of the predicted segmentation results to alleviate the domain shift. Then, direct entropy-minimization optimization is applied to the unlabeled target domain data to improve the credibility of the prediction segmentation maps. To enhance the prediction consistency of the target domain data, we augment the target domain dataset through the Fourier transform by replacing the low-frequency part in the target images with that of the source images. Then, a semantic consistency constraint is imposed on the raw images and augmented images of the target domain to improve the prediction consistency of the segmentation model, thereby further narrowing the discrepancy between the source and target domains. Experiments on several public retinal fundus image datasets prove the superiority of MeFDA compared with state-of-the-art methods, and the ablation study analyzes the importance of the different proposed components.

When a target is moving at high-speed, its high-resolution range profile (HRRP) will be stretched by the high-order phase error caused by the high velocity. In this case, the inverse synthetic aperture radar (ISAR) image would be seriously blurred. To obtain a well-focused ISAR image, the phase error induced by target velocity should be compensated. This article exploits the variation continuity of a high-speed moving target's velocity and proposes a noise-robust high-speed motion compensation algorithm for ISAR imaging. The target's velocity within a coherent processing interval (CPI) is modeled as a high-order polynomial based on which a parametric high-speed motion compensation signal model is developed. The entropy of the ISAR image after high-speed motion compensation is treated as an evaluation metric, and a parametric minimum entropy optimization model is established to estimate the velocity and compensate it simultaneously. A gradient-based solver of this optimization is then adopted to iteratively find the optimal solution. Finally, the high-order phase error caused by the target's high-speed motion can be iteratively compensated, and a well-focused ISAR image can be obtained. Extensive simulation experiments have verified the noise robustness and effectiveness of the proposed algorithm. [ABSTRACT FROM AUTHOR]

We consider asset pricing models in which the SDF can be factorized into an observable component and a potentially unobservable one. Using a relative entropy minimization approach, we nonparametrically estimate the SDF and its components. Empirically, we find the SDF has a business-cycle pattern and significant correlations with market crashes and the Fama-French factors. Moreover, we derive novel bounds for the SDF that are tighter and have higher information content than existing ones. We show that commonly used consumption-based SDFs correlate poorly with the estimated one, require high risk aversion to satisfy the bounds and understate market crash risk. [ABSTRACT FROM AUTHOR]

The article is devoted to the approximate solutions of the Fredholm integral equations of the second kind with the weak singular kernel that can have additional singularity in the numerator. We describe two problems that lead to such equations. They are the problem of minimization of small deviations and the entropy minimization problem. Both of them appear when considering a dynamical system involving a mixed fractional Brownian motion. In order to apply well-known numerical methods for weakly singular kernels, we build the continuous approximation of the solution of an integral equation with the kernel containing additional singularity by the solutions of the integral equations whose kernels are weakly singular, but the numerator is continuous. We prove that the approximated solutions tend to the solution of the original equation. We demonstrate numerically how our methods work being applied to our specific integral equations. [ABSTRACT FROM AUTHOR]

The wings of an airplane vibrate when it is nonstationary. When an airplane is observed using a microwave photonics-based ultrahigh-resolution radar, and this vibration will cause defocusing of the wings. To address this problem, we propose a vibration-parameter estimation method for airplane wings based on microwave-photonics ultrahigh-resolution radar. In this method, we first coarsely separate images of the wings and body of the airplane, and estimate the Light-Of-Sight (LOS) of the radar from a focused and scaled image of the airplane body. Next, we apply sub-aperture imaging to the wings and extract the range and Doppler curves from a sequence of sub-aperture images. By combining the range and Doppler curves with the LOS, we can obtain a preliminary estimation of the vibration parameters. Finally, by Modifying the Polar Format Algorithm (MPFA) and constructing an optimization function that minimizes image entropy, we can obtain accurate vibration parameters. This novel modified polar format algorithm can be applied to complex motion targets, such as an airplane with vibrating wings, swinging ships to effectively decouple range and cross-range dimensional coupling. Experimental results using both simulated and measured data confirm the validity and practicality of the proposed algorithm.

We derive a multi-species BGK model with velocity-dependent collision frequency for a non-reactive, multi-component gas mixture. The model is derived by minimizing a weighted entropy under the constraint that the number of particles of each species, total momentum, and total energy are conserved. We prove that this minimization problem admits a unique solution for very general collision frequencies. Moreover, we prove that the model satisfies an H-Theorem and characterize the form of equilibrium. [ABSTRACT FROM AUTHOR]

Copyright of Journal of Zhengzhou University (Natural Science Edition) is the property of Journal of Zhengzhou University (Natural Science Edition) Editorial Office and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Given a collection of probability distributions p1, . . . , pm, the minimum entropy coupling is the coupling X1, . . . , Xm (Xi ∼ pi) with the smallest entropy H(X1, . . . , Xm). While this problem is known to be NP-hard, we present an efficient algorithm for computing a coupling with entropy within 2 bits from the optimal value. More precisely, we construct a coupling with entropy within 2 bits from the entropy of the greatest lower bound of p1, . . . , pm with respect to majorization. This construction is also valid when the collection of distributions is infinite, and when the supports of the distributions are infinite. Potential applications of our results include random number generation, entropic causal inference, and functional representation of random variables. [ABSTRACT FROM AUTHOR]

Deep learning models are sensitive to domain shift phenomena. A model trained on images from one domain cannot generalise well when tested on images from a different domain, despite capturing similar anatomical structures. It is mainly because the data distribution between the two domains is different. Moreover, creating annotation for every new modality is a tedious and time-consuming task, which also suffers from high inter- and intra- observer variability. Unsupervised domain adaptation (UDA) methods intend to reduce the gap between source and target domains by leveraging source domain labelled data to generate labels for the target domain. However, current state-of-the-art (SOTA) UDA methods demonstrate degraded performance when there is insufficient data in source and target domains. In this paper, we present a novel UDA method for multi-modal cardiac image segmentation. The proposed method is based on adversarial learning and adapts network features between source and target domain in different spaces. The paper introduces an end-to-end framework that integrates: a) entropy minimization, b) output feature space alignment and c) a novel point-cloud shape adaptation based on the latent features learned by the segmentation model. We validated our method on two cardiac datasets by adapting from the annotated source domain, bSSFP-MRI (balanced Steady-State Free Procession-MRI), to the unannotated target domain, LGE-MRI (Late-gadolinium enhance-MRI), for the multi-sequence dataset; and from MRI (source) to CT (target) for the cross-modality dataset. The results highlighted that by enforcing adversarial learning in different parts of the network, the proposed method delivered promising performance, compared to other SOTA methods. [ABSTRACT FROM AUTHOR]

CO2 emission control is one of the most vital parts of environment management. China owns the largest CO2 emission in the world. For the sake of clarifying China's emission sharing responsibilities and set emission reduction targets, a considerable number of scholars have worked to project China's embodied CO2 emission. Single Regional Input–output (SRIO) model is widely used for investigating CO2 emission issues. Considering the ubiquitous time lag of input–output data, entropy optimization model is introduced to estimate SRIO tables. However, the uncertainty in the corresponding model parameters necessarily has a serious impact on the estimation results. To consider the impact of uncertainties, we introduce robust optimization into entropy minimization model for SRIO table estimation. Based on three different uncertainty sets, we constructed three robust entropy minimization models to construct 2016 China's SRIO tables and calculate China's embodied CO2 emission based on those tables. The estimation results show that the model based on the ball uncertainty set has the best performance with less uncertainty, while the model based on the budgeted uncertainty set performances more 'robust' facing greater uncertainty, which means its performance is less volatile at different levels of uncertainty. Moreover, the embodied carbon emission is predicted to reach 9632.57 Mt CO2. The top emitter is the sector of supply of electricity, heating and water, which accounts for more than 40% of total CO2 emission. [ABSTRACT FROM AUTHOR]