Dense Architecture Research Articles

Balanced spatio-spectral feature extraction for hyperspectral and multispectral image fusion

Although remote sensing sensors acquire hyperspectral images (HSI) with outstanding spectral resolution, hardware limitations prevent them from obtaining HSI with superior spatial resolution. Having high-resolution multispectral images (MSI) from the same scene has sparked the concept of HSI-MSI fusion to achieve high-resolution HSI. Recently, deep learning algorithms have shown excellent performance in HSI-MSI fusion. Deep learning algorithms can automatically extract the priors latent in the input data and reconstruct the output image. However, their performance depends on the training data’s inherent characteristics. HSI volume is generally significantly larger than MSI volume. While the high-resolution MSI provides richer spatial information, its contribution to fusion is lower due to its smaller volume. To address this volume discrepancy, a deep convolutional neural network named BASFE is proposed. This network employs a two-branch structure that balances the extracted spatio-spectral features from both HSI and MSI input images, enabling effective feature fusion. Therefore, better efficiency can be achieved by using simpler structures. Experiments showed that balancing the data improves the model’s performance significantly. Furthermore, to improve efficiency and reduce complexity and computational burden, the spatio-spectral feature extraction residual modules are embedded in a dense architecture to jointly extract these features. Comprehensive experiments on five challenging datasets show that BASFE surpasses state-of-the-art algorithms. Compared to the first- and second-best competing methods, the proposed method achieves an average improvement of 1.58 dB and 3.68 dB in peak signal-to-noise ratio (PSNR) across the five studied datasets. The source code of BASFE is available in https://github.com/rajaei-arash/BASFE_hyperspectral_fusion.git.

Pressure‐Induced Dense and Robust Ge Architecture for Superior Volumetric Lithium Storage

AbstractThe germanium (Ge) anode attains wide attention in lithium‐ion batteries because of its high theoretical volumetric capacity (8646 mAh cm−3). However, the huge volume expansion (≈230%) results in its poor electrochemical performances. The strategies reported in the literature to solve the issue often cause a low packing density, lowering the volumetric capacity. Here, a pressure‐induced route is proposed to fabricate a Ge architecture, in which nano‐sized Ge (≈15 nm) is encapsulated by robust TiO2 and highly conductive carbon, which offer the advantages of a low stress–strain characteristic, low volume expansion in thickness change, high electrical conductivity (463.2 S m−1), high Li‐ion diffusion coefficient (9.55 × 10−9–8.51 × 10−12 cm2 s−1), and high tapping density (1.79 g cm−3). As a result, the dense architecture obtains outstanding volumetric capacities of 3559.8 mAh cm−3 at 0.1 A g−1 and 2628.2 mAh cm−3 at 20 A g−1, along with excellent cycling life over 5000 cycles at 10 A g−1. Remarkably, the full cell achieves a high volumetric energy density of 1760.1 Wh L−1, along with impressive fast‐charging performances and long cycling life. This work provides a new synthesis strategy and deep insight into the design of high‐volumetric capacity alloy‐based lithium‐ion‐battery anodes.

Colon and lung cancer classification from multi-modal images using resilient and efficient neural network architectures

Automatic classification of colon and lung cancer images is crucial for early detection and accurate diagnostics. However, there is room for improvement to enhance accuracy, ensuring better diagnostic precision. This study introduces two novel dense architectures (D1 and D2) and emphasizes their effectiveness in classifying colon and lung cancer from diverse images. It also highlights their resilience, efficiency, and superior performance across multiple datasets. These architectures were tested on various types of datasets, including NCT-CRC-HE-100K (set of 100,000 non-overlapping image patches from hematoxylin and eosin (H&E) stained histological images of human colorectal cancer (CRC) and normal tissue), CRC-VAL-HE-7K (set of 7180 image patches from N = 50 patients with colorectal adenocarcinoma, no overlap with patients in NCT-CRC-HE-100K), LC25000 (Lung and Colon Cancer Histopathological Image), and IQ-OTHNCCD (Iraq-Oncology Teaching Hospital/National Center for Cancer Diseases), showcasing their effectiveness in classifying colon and lung cancers from histopathological and Computed Tomography (CT) scan images. This underscores the multi-modal image classification capability of the proposed models. Moreover, the study addresses imbalanced datasets, particularly in CRC-VAL-HE-7K and IQ-OTHNCCD, with a specific focus on model resilience and robustness. To assess overall performance, the study conducted experiments in different scenarios. The D1 model achieved an impressive 99.80 % accuracy on the NCT-CRC-HE-100K dataset, with a Jaccard Index (J) of 0.8371, a Matthew's Correlation Coefficient (MCC) of 0.9073, a Cohen's Kappa (Kp) of 0.9057, and a Critical Success Index (CSI) of 0.8213. When subjected to 10-fold cross-validation on LC25000, the D1 model averaged (avg) 99.96 % accuracy (avg J, MCC, Kp, and CSI of 0.9993, 0.9987, 0.9853, and 0.9990), surpassing recent reported performances. Furthermore, the ensemble of D1 and D2 reached 93 % accuracy (J, MCC, Kp, and CSI of 0.7556, 0.8839, 0.8796, and 0.7140) on the IQ-OTHNCCD dataset, exceeding recent benchmarks and aligning with other reported results. Efficiency evaluations were conducted in various scenarios. For instance, training on only 10 % of LC25000 resulted in high accuracy rates of 99.19 % (J, MCC, Kp, and CSI of 0.9840, 0.9898, 0.9898, and 0.9837) (D1) and 99.30 % (J, MCC, Kp, and CSI of 0.9863, 0.9913, 0.9913, and 0.9861) (D2). In NCT-CRC-HE-100K, D2 achieved an impressive 99.53 % accuracy (J, MCC, Kp, and CSI of 0.9906, 0.9946, 0.9946, and 0.9906) with training on only 30 % of the dataset and testing on the remaining 70 %. When tested on CRC-VAL-HE-7K, D1 and D2 achieved 95 % accuracy (J, MCC, Kp, and CSI of 0.8845, 0.9455, 0.9452, and 0.8745) and 96 % accuracy (J, MCC, Kp, and CSI of 0.8926, 0.9504, 0.9503, and 0.8798), respectively, outperforming previously reported results and aligning closely with others. Lastly, training D2 on just 10 % of NCT-CRC-HE-100K and testing on CRC-VAL-HE-7K resulted in significant outperformance of InceptionV3, Xception, and DenseNet201 benchmarks, achieving an accuracy rate of 82.98 % (J, MCC, Kp, and CSI of 0.7227, 0.8095, 0.8081, and 0.6671). Finally, using explainable AI algorithms such as Grad-CAM, Grad-CAM++, Score-CAM, and Faster Score-CAM, along with their emphasized versions, we visualized the features from the last layer of DenseNet201 for histopathological as well as CT-scan image samples. The proposed dense models, with their multi-modality, robustness, and efficiency in cancer image classification, hold the promise of significant advancements in medical diagnostics. They have the potential to revolutionize early cancer detection and improve healthcare accessibility worldwide.

A facile method to fabricate versatile keratin cryogels for tissue engineering applications

Human hair keratin (HHK) has been extensively explored as a biomaterial for soft tissue regeneration due to their excellent bioactivity and biocompatibility. The possibility to fabricate HHK into three-dimensional (3D) hydrogels with physical properties resembling soft tissues has been well demonstrated. However, conventional keratin hydrogels often exhibit a dense architecture that could hinder cell filtration. In the present study, HHK-based cryogels were fabricated using a freeze-thaw (FT) method, where oxidized dopamine (ODA) was employed to covalently crosslink thiol/amine rich-keratin molecules at sub-zero temperatures. The obtained HHK-ODA cryogels have micron-sized pores ranging between 100 and 200 μm and mechanical properties that can be tuned by varying the crosslinking density between ODA and HHK. Through optimization of the weight content of ODA and the number of FT cycles, the compressive strengths and stiffnesses of these cryogels achieved 15-fold increments from ∼1.5 kPa to ∼22 kPa and ∼300 Pa to ∼5000 Pa, respectively. The HHK-ODA cryogels competently supported human dermal fibroblast spreading and proliferation. Overall, this study exhibited a facile method to fabricate mechanically superior keratin-based cryogels with cell compatible microarchitecture, circumventing the need for complicated chemical modifications and the use of cytotoxic crosslinkers.

Deep learning for sentiment analysis on IMDB movie reviews using N-gram features

In the rapidly evolving digital landscape, the synergy of deep learning techniques and abundant datasets has opened new frontiers in various domains. This research delves into the film industry, specifically harnessing the potential of the International Movie DataBase (IMDB) dataset for sentiment analysis. Through a deep learning paradigm, we embark on sentiment classification of movie reviews, discerning between positive and negative sentiments. By navigating data preprocessing and N-gram feature extraction, we engineer a deep learning model comprising embedding, global average pooling, and multi-layer dense architectures. The experimental results underscore the model's prowess in sentiment analysis, emphasizing its capacity to empower informed decision-making within the film industry.

Actinin-4 controls survival signaling in B cells by limiting the lateral mobility of B-cell antigen receptors.

The structure and dynamics of F-actin networks in the cortical area of B cells control the signal efficiency of B-cell antigen receptors (BCRs). Although antigen-induced signaling has been studied extensively, the role of cortical F-actin in antigen-independent tonic BCR signaling is less well understood. Because these signals are essential for the survival of B cells and are consequently exploited by several B-cell lymphomas, we assessed how the cortical F-actin structure influences tonic BCR signal transduction. We employed genetic variants of a primary cell-like B-cell line that can be rendered quiescent to show that cross-linking of actin filaments by α-actinin-4 (ACTN4), but not ACTN1, is required to preserve the dense architecture of F-actin in the cortical area of B cells. The reduced cortical F-actin density in the absence of ACTN4 resulted in increased lateral BCR diffusion. Surprisingly, this was associated with reduced tonic activation of BCR-proximal effector proteins, extracellular signal-regulated kinase, and pro-survival pathways. Accordingly, ACTN4-deficient B-cell lines and primary human B cells exhibit augmented apoptosis. Hence, our findings reveal that cortical F-actin architecture regulates antigen-independent tonic BCR survival signals in human B cells.

Facile Bond Exchanging Strategy for Engineering Wet Adhesion and Antioxidant/Antibacterial Thin Layer over a Dynamic Hydrogel via the Carbon Dots Derived from Tannic Acid/ε-Polylysine.

Adhesive hydrogels, playing an essential role in stretchable electronics, soft robotics, tissue engineering, and so forth, upon functioning often need to adhere to various substrates in wet conditions and simultaneously exhibit antibacterial/antioxidant properties while possessing the intrinsic stretchability and elasticity of the hydrogel network intact. Therefore, simple approaches to conveniently access adhesive hydrogels with multifunctional surfaces are being pursued. Herein, a facile strategy has been proposed to construct multifunctional adhesive hydrogels via surface engineering of a multifunctional carbon dot (CD)-decorated polymeric thin layer by dynamic bond exchange. By this strategy, a double cross-linked network hydrogel of polyacrylamide (PAM) and oxidized dextran (ODA) was engineered with a unique dense layer over the Schiff base hydrogel matrix by aqueous solution immersion of PA-120, versatile CDs derived from tannic acid (TA) and ε-polylysine (PL). Without any additional agents, the PA-120 CDs with residual polyphenolic/catechol and amine moieties were incorporated into the surface structure of the hydrogel network by the combined action of the Schiff base and hydrogen bonds to form a dense surface layer that can exhibit high wet adhesive performance via the amine-polyphenol/catechol pair. The armor-like dense architecture also endowed hydrogels with considerably enhanced tensile/compression properties and excellent antioxidant/antibacterial abilities. Besides, the single-sided modified Janus hydrogel and completely surface-modified hydrogel can be flexibly developed through this approach. This strategy will provide new insights into the preparation and application of surface-modified hydrogels featuring multiple functions and tunable interfacial properties.

Tuning the nucleation kinetics of phosphate chemical conversion coating on Mg-Gd-Y-Zr magnesium alloy: The effect of pretreatment and organic additive

In order to obtain a more protective phosphate conversion coating with a denser architecture, the nucleation kinetics of phosphate chemical conversion coating on Mg-Gd-Y-Zr magnesium alloy was tuned in this work. A pretreatment process was proposed and organic additives were incorporated, which aims at increasing the ionic produce (Jsp) at the interface for increasing σ, and decreasing the critical ionic product (JC,sp), respectively. Results prove that the pretreatment of bare alloys in a phosphate bath could increase the ion products of MgHPO4 / MnHPO4. The addition of benzalkonium chloride could neutralize the charges of crystals, and in turn promote the nucleation kinetics. A denser and more protective conversion coating could consequently be obtained.

Neural network-based emulation of interstellar medium models

Context. The interpretation of observations of atomic and molecular tracers in the galactic and extragalactic interstellar medium (ISM) requires comparisons with state-of-the-art astrophysical models to infer some physical conditions. Usually, ISM models are too timeconsuming for such inference procedures, as they call for numerous model evaluations. As a result, they are often replaced by an interpolation of a grid of precomputed models. Aims. We propose a new general method to derive faster, lighter, and more accurate approximations of the model from a grid of precomputed models for use in inference procedures. Methods. These emulators are defined with artificial neural networks (ANNs) with adapted architectures and are fitted using regression strategies instead of interpolation methods. The specificities inherent in ISM models need to be addressed to design and train adequate ANNs. Indeed, such models often predict numerous observables (e.g., line intensities) from just a few input physical parameters and can yield outliers due to numerical instabilities or physical bistabilities and multistabilities. We propose applying five strategies to address these characteristics: (1) an outlier removal procedure; (2) a clustering method that yields homogeneous subsets of lines that are simpler to predict with different ANNs; (3) a dimension reduction technique that enables us to adequately size the network architecture; (4) the physical inputs are augmented with a polynomial transform to ease the learning of nonlinearities; and (5) a dense architecture to ease the learning of simpler relations between line intensities and physical parameters. Results. We compare the proposed ANNs with four standard classes of interpolation methods, nearest-neighbor, linear, spline, and radial basis function (RBF), to emulate a representative ISM numerical model known as the Meudon PDR code. Combinations of the proposed strategies produce networks that outperform all interpolation methods in terms of accuracy by a factor of 2 in terms of the average error (reaching 4.5% on the Meudon PDR code) and a factor of 3 for the worst-case errors (33%). These networks are also 1000 times faster than accurate interpolation methods and require ten to forty times less memory. Conclusions. This work will enable efficient inferences on wide-field multiline observations of the ISM.

Efficient Multi-Task CNN for Face and Facial Expression Recognition Using Residual and Dense Architectures for Application in Monitoring Online Learning

Efficient Multi-Task CNN for Face and Facial Expression Recognition Using Residual and Dense Architectures for Application in Monitoring Online Learning

Deep-Learning Approach for the Atomic Configuration Interaction Problem on Large Basis Sets.

High-precision atomic structure calculations require accurate modeling of electronic correlations typically addressed via the configuration interaction (CI) problem on a multiconfiguration wave function expansion. The latter can easily become challenging or infeasibly large even for advanced supercomputers. Here, we develop a deep-learning approach which allows us to preselect the most relevant configurations out of large CI basis sets until the targeted energy precision is achieved. The large CI computation is thereby replaced by a series of smaller ones performed on an iteratively expanding basis subset managed by a neural network. While dense architectures as used in quantum chemistry fail, we show that a convolutional neural network naturally accounts for the physical structure of the basis set and allows for robust and accurate CI calculations. The method was benchmarked on basis sets of moderate size allowing for the direct CI calculation, and further demonstrated on prohibitively large sets where the direct computation is not possible.

Physics-model-free heat-electricity energy management of multiple microgrids based on surrogate model-enabled multi-agent deep reinforcement learning

Reliability and cost-effectiveness in the operation of the multiple microgrid (MMG) system depend on the skillful management of its energy resources. Traditional energy management approaches are physics-model-based, which rely on the precise system parameters (e.g., line parameters) of the electricity and heat network. It is difficult to establish the precise system parameters in practice because they depend on a variety of factors. In this context, this paper proposes a physics-model-free control framework for the energy management of MMGs with heat-electricity energy, consisting of proposed surrogate model and multi-agent deep reinforcement learning (MADRL) approach. An important step is to use historical data to train a surrogate model in supervised manner that can imitate the realistic power and thermal flow calculations. Meanwhile, the energy management problem is reformulated as a Markov game. It is solved by the proposed MADRL-based approach by modeling each MG controller as an agent with a specific objective. The historical trajectories representation, parameter space technology, and deep dense architecture in reinforcement learning are introduced in MADRL to overcome the negative impact brought by the time series data from the input state on the decision-making process and construct an efficient exploration mechanism to overcome inefficient optimization of the MMG system in a multi-agent setting. During training period of MADRL, the trained surrogate models integrate into the environment of the MADRL, which can develop optimal energy management strategy based on the continuous interaction with the surrogate models. The proposed surrogate model enabled MADRL approach can reduce the reliance on precise physical systems and prevent having an impact on the real system while being trained involving trial and error process. Simulation results demonstrate the effectiveness of the proposed control framework.

Embedded 3D Printing of Architected Ceramics via Microwave-Activated Polymerization.

Light- and ink-based 3D printing methods have vastly expanded the design space and geometric complexity of architected ceramics. However, light-based methods are typically confined to a relatively narrow range of preceramic and particle-laden resins, while ink-based methods are limited in geometric complexity due to layerwise assembly. Here, embedded 3D printing is combined with microwave-activated curing to generate architected ceramics with spatially controlled composition in freeform shapes. Aqueous colloidal inks are printed within a support matrix, rapidly cured via microwave-activated polymerization, and subsequently dried and sintered into dense architectures composed of one or more oxide materials. Thisintegrated manufacturing method opens new avenues for the design and fabrication of complex ceramic architectures with programmed composition, density, and form for myriad applications.

The perturbational map of low frequency repetitive transcranial magnetic stimulation of primary motor cortex in movement disorders

BackgroundRepetitive transcranial magnetic stimulation (rTMS) is applied to the primary motor cortex (M1) for the treatment of different movement disorders like Writer's Cramp (WC), Essential tremor (ET), and Spinocerebellar ataxia (SCA). However, the benefits vary, ranging from no effect to significant improvement in tremor. The variation in the benefits obtained from rTMS might be due to the difference in the spread of the stimuli across the brain areas associated with tremor. The spread of stimuli can be evaluated by combining rTMS with resting state functional magnetic resonance imaging (rsfMRI). AimTo determine the spread of low frequency rTMS for WC, ET and SCA after stimulation of M1. MethodThe rsfMRI was collected from the participants with WC (n = 27), ET (n = 30) and SCA (n = 28) at two time points, i.e., before and after the delivery of 1 Hz rTMS. Two measures from dynamic systems theory were calculated to understand how the system interacts with exogenously applied input (rTMS), namely entropy and frustration. While entropy measures the disorder from the rsfMRI time series, frustration quantifies it by assessing the change in sign (positive to negative, and vice-versa) of the functional connections. The two quantities outlined the spread of perturbation due to rTMS. ResultFor WC, a dense architecture of functional connections facilitated the spread across the bilateral areas of the five regions namely- the frontal cortex (frontal-Mid and Sup), motor cortex (supplemental motor area, precentral- and postcentral gyrus), parietal cortex (precuneus), subcortical regions (caudate, thalamus, posterior and mid-cingulate cortex) and the cerebellum (Crus-I and II, Cerebellum III and VI, and Vermis VI). For ET, though the areas belong to the above-mentioned 5 regions, the spread was narrower (i.e., lesser areas in the region). For SCA, the sparse connection between the areas led to minimal spread with no propagation of rTMS perturbation to the cerebellar regions. Interestingly for the three disorders, rTMS reduced the disorderliness (frustration and entropy) of the neural circuit (motor, subcortical and cerebellar regions) but increased the disorderliness in the default mode network (frontal and parietal regions). ConclusionThe spread of perturbation (due to TMS on M1) in the functional circuit varies in the three movement disorders. The current rTMS protocol achieves a wider spread for WC but not for ET and SCA. Findings from the present study suggest that pathology-specific stimulation protocols are required in movement disorders.

Deep Dense Network-Based Curriculum Reinforcement Learning for High-Speed Overtaking

Reinforcement learning methods have promising applications in autonomous vehicles, such as lane keeping, adaptive cruise control, and overtaking. However, long-time sequential decision making in complicated scenarios, such as continuous high-speed overtaking, remains challenge. One such challenge is that current neural networks in reinforcement learning employ a shallow neural network, which has limitations in feature extractions. To address this, we propose a deep actor–critic network framework using multilayer dense-connection networks to obtain efficient extracted features. To be specific, we employ a deep dense architecture in the framework of soft actor–critic, which has multiple hidden layers serving as feature extractions. To tackle the vanishing or exploding gradient problem, we reuse features and design shortcut connections, as well as feature-concatenating operations between hidden layers. Meanwhile, when learning multitask high-speed overtaking as one whole progression, the agent has to learn driving, lane changing, and avoiding other vehicles at the same time. Therefore, another challenge is that sparse rewards worsen the speed of convergence. We design a task-level sequencing curriculum reinforcement learning method to tackle the reward sparseness. It breaks down a complex task into two successive subtasks: the first aims to drive as fast as possible in a single-vehicle model; and the second aims to overtake other vehicles without any collisions as soon as possible according to the knowledge gained from the first task in a multivehicle model. Finally, we evaluate our method using a racing car simulator with different tracks. In a comparison using the standard actor–critic algorithm in three different multivehicle scenarios, the best median results show that our deep actor–critic network framework reduces overtaking distance and overtaking time by up to 4.9% and 8.7%, respectively.

IoT based optical coherence tomography retinal images classification using OCT Deep Net2

Machine learning algorithms gains prominence in health care sectors for disease diagnosis, classification and prediction. Deep learning architecture gains prominence in real time applications. This research work proposes a novel deep learning architecture, OCT Deep Net2 for the classification of optical coherence tomography images. Four class disease classification was performed in this research work and the proposed deep learning framework OCT Deep Net2 is an extension of OCT Deep Net1 comprising of 30 layers. The OCT Deep Net2 comprises of 50 layers and is termed as dense architecture, comprises of three recurrent modules. The performance validation reveals the efficiency of the OCT Deep Net2 architecture in terms of the performance metrics. Robust results were produced for batch size of 32 and 100 epochs with an accuracy of 98%. An IoT based system was implemented using Raspberry Pi B+ processor, the OCT Deep Net 2 algorithm was written in Python and executed on Google Colab.

Multi-centric, Marsh-based Urbanism at the early Mesopotamian city of Lagash (Tell al-Hiba, Iraq)

• Large-scale remote sensing data show urban structure at 3rd millennium BCE Lagash. • Lagash’s urban form challenges the standard model of the world’s earliest cities. • >300 ha of architecture are clustered in bounded sectors separated by empty spaces. • City sectors perhaps began as marsh islands; discontinuous areas shaped city life. • Early southern Mesopotamian cities may have frequently been spatially multi-centric. Leveraging a suite of remote sensing technologies deployed over large areas, this paper presents results that challenge long-held ideas about the origin and development of the world’s oldest urban centers, in southern Iraq. The standard model of third millennium BCE Mesopotamian cities presents them as nuclear, compact settlements set within an irrigated agricultural hinterland, expanding continuously from a monumental religious complex. This reconstruction holds enormous influence in the comparative global study of early urbanism. UAV photos and magnetic gradiometry data captured at Lagash (Tell al-Hiba) show dense architecture and related paleoenvironmental features over c. 300 ha, revealing a city that does not conform to the standard model. Early Dynastic Lagash (2900–2350 BCE) was composed of spatially discrete sectors bounded by multiple surrounding walls and/or watercourses and separated by open spaces. The evidence is suggestive of a marshy or watery local environment, and the city sectors may have originated as marsh islands. The discontinuous, walled nature of inhabited areas would have had social and logistical ramifications for city inhabitants. A number of contemporary sites are characterized by multiple archaeological mounds, suggesting that early southern Mesopotamian cities may have frequently been spatially multi-centric.

Transcriptomics reveal the antibiofilm mechanism of NaCl combined with citral against Vibrio parahaemolyticus.

Vibrio parahaemolyticus (VP) as a prominent foodborne pathogen in seafood generally adheres to various surfaces and forms biofilms in the processing of aquatic products. The study aimed to elucidate the inhibitory efficacy and potential mechanism of salinity (NaCl) or it combined with citral against the biofilm formation of VP. Three VP strains formed the most biofilm at 1.0% NaCl, and their biofilms gradually declined with the increase of NaCl concentration. Compared with 1% NaCl, applying 3% and 5% NaCl or NaCl in combination with citral at 10-40µg/mL significantly reduced biofilm biomass, cellular activity, and viable cells, as well as extracellular polymeric substances (EPS) and cell surface hydrophobicity. Sparser and thinner VP biofilm with large dead cells were observed under the combined treatment, in contrast to the dense architectures of biofilm formed at 1% NaCl. Although VP exhibited the strongest swimming and swarming ability at 3% NaCl, the two motilities were both significantly reduced by citral for all three salinities. Transcriptomic profiling revealed that, compared with 1% NaCl (Con), the two treatments consisting of 3% NaCl (Sal3) and it combined with 40µg/mL citral (Com) drastically altered gene expression patterns in VP biofilm cells, resulting in 1196 and 1304 differentially expressed genes, respectively. The treatment of Com group altered the transcription of various genes related to chemotaxis, flagellar assembly, EPS synthesis, LuxS and CqsA-mediated quorum sensing, and c-di-GMP, which might interfere with biofilm development of VP. Our findings provided novel insights into the combined regulatory mechanism of high salinity and citral for antibiofilm formation in VP. KEY POINTS: • High salinity enhanced the antibiofilm efficacy of citral against Vibrio parahaemolyticus. • Combined treatment downregulated the expression of exopolysaccharide synthesis genes. • A total of 3% NaCl and combined treatments interfered with signaling pathways of QS and c-di-GMP.

A Multiagent Deep Reinforcement Learning Based Approach for the Optimization of Transformer Life Using Coordinated Electric Vehicles

The uncertainties of charging behavior of electric vehicle (EV) owners have a negative impact on the loss of life (LOL) of distribution transformer. This article proposes a decentralized EV charging framework for optimization of the LOL of distribution transformer considering the dissatisfactions of EV owners. Specifically, long-short-term memory (LSTM) neural network is first utilized to capture the uncertainties caused by the load demand and electricity price. After that, each EV is modeled as an intelligent agent and a multiagent deep reinforcement learning approach is applied to solve the coordinated charging problem based on the forecasting information by the LSTM network. All the agents are trained in a centralized manner to develop coordinated control strategies while informing decisions based on local information when finishing the training process. The proposed approach can achieve coordinated charging management of EVs based on local information, which helps preserve the privacy of EV owners, reduce the cost induced by the deployment of communication devices, and avoid single-point failure. In addition, the parameter space noise and deep dense architecture in reinforcement learning are introduced to overcome premature convergence, training instability, and inefficiency due to the large action space of multiagent scenario. Comparative tests are carried out among several benchmarks utilizing real-world data to illustrate the effectiveness of the proposed approach.

A Robust Framework for Object Detection in a Traffic Surveillance System

Object recognition is the technique of specifying the location of various objects in images or videos. There exist numerous algorithms for the recognition of objects such as R-CNN, Fast R-CNN, Faster R-CNN, HOG, R-FCN, SSD, SSP-net, SVM, CNN, YOLO, etc., based on the techniques of machine learning and deep learning. Although these models have been employed for various types of object detection applications, however, tiny object detection faces the challenge of low precision. It is essential to develop a lightweight and robust model for object detection that can detect tiny objects with high precision. In this study, we suggest an enhanced YOLOv2 (You Only Look Once version 2) algorithm for object detection, i.e., vehicle detection and recognition in surveillance videos. We modified the base network of the YOLOv2 by reducing the number of parameters and replacing it with DenseNet. We employed the DenseNet-201 technique for feature extraction in our improved model that extracts the most representative features from the images. Moreover, our proposed model is more compact due to the dense architecture of the base network. We utilized DenseNet-201 as a base network due to the direct connection among all layers, which helps to extract a valuable information from the very first layer and pass it to the final layer. The dataset gathered from the Kaggle and KITTI was used for the training of the proposed model, and we cross-validated the performance using MS COCO and Pascal VOC datasets. To assess the efficacy of the proposed model, we utilized extensive experimentation, which demonstrates that our algorithm beats existing vehicle detection approaches, with an average precision of 97.51%.