Core Module Research Articles

Aerial imagery-based tobacco plant counting framework for efficient crop emergence estimation

Crop emergence estimation at early crop growth stages is becoming increasingly important for the long-term sustainability of natural resources. It helps farmers and agricultural stakeholders in the efficient allocation of resources like water, pesticides, and fertilizers. It can be used to estimate the yield and seed quality, identify the region of potential yield losses, and make future agriculture plans. These future agriculture plans can play a crucial role in ensuring maximum crop population and yield while utilizing the same limited land and natural resources. Most of the existing plant counting frameworks require offline processing of images with computationally expensive algorithms including the structure for motion and multiview stereo to develop an orthomosaic. This study proposed a tobacco plant counting framework that directly estimates counts from aerial images and has the potential for real-time applicability. It consists of three core modules: overlap detection, plant detection, and plant counting. The overlap detection module replaces the need for computationally expensive orthomosaic formation to avoid counting repetition by overlap masking based on only visual cues. Three different methods are evaluated as core modules for finding an optimal solution for plant counting based on time complexity and accuracy. In the first method after overlap detection, semantic segmentation with U-NET is employed as a plant detection module. For plant counting, we count the connected pixels classified as plants to estimate the crop count. In the second method after overlap detection, object detection using YOLOv7 is utilized as a plant detection module followed by simply counting each detected plant. In the third method, we utilize YOLOv7 for object detection, similar to the second method. However, we introduce the SORT (Simple Online and Realtime Tracking) algorithm for object tracking. This object tracking replaces the overlap detection module making it a real-time applicable method. For plant counting, we assess the number of tracked plants. The proposed algorithm is evaluated on two distinct tobacco fields. The high-resolution aerial data is collected from tobacco fields near Peshawar, Pakistan, and is human-labelled. The first and second methods show average F1 scores of 0.947 and 0.9667, respectively, whereas the third method has the potential for real-time applicability with an average F1 score of 0.967.

Application of a user experience design approach for an EHR-based clinical decision support system.

We applied a user experience (UX) design approach to clinical decision support (CDS) tool development for the specific use case of pediatric asthma. Our objective was to understand physicians' workflows, decision-making processes, barriers (ie, pain points), and facilitators to increase usability of the tool. We used a mixed-methods approach with semi-structured interviews and surveys. The coded interviews were synthesized into physician-user journey maps (ie, visualization of a process to accomplish goals) and personas (ie, user types). Interviews were conducted via video. We developed physician journey maps and user personas informed by their goals, systems interactions, and experiences with pediatric asthma management. The physician end-user personas identified were: efficiency, relationship, and learning. Features of a potential asthma CDS tool sought varied by physician practice type and persona. It was important to the physician end-user that the asthma CDS tool demonstrate value by lowering workflow friction (ie, difficulty or obstacles), improving the environment surrounding physicians and patients, and using it as a teaching tool. Customizability versus standardization were important considerations for uptake. Different values and motivations of physicians influence their use and interaction with the EHR and CDS tools. These different perspectives can be captured by applying a UX design approach to the development process. For example, with the importance of customizability, one approach may be to build a core module with variations depending on end-user preference. A UX approach can drive design to help understand physician-users and meet their needs; ultimately with the goal of increased uptake.

A simplified geospace model for satellite design

This study introduced a simplified, integrated computational model to estimate the drag coefficient and surface charging within the plasma wake of Low Earth Orbit (LEO) spacecraft influenced by various solar activities. The model constituted four core modules: a solver for the solar wind-magnetosphere interaction, a magnetosphere-ionosphere coupling for electric field mapping, an ionization model for the F-layer of the ionosphere, and an estimation model for excess satellite drag and plasma wake size. The initial module adopted a simplified 1D ideal Magnetohydrodynamic (MHD) model while incorporating magnetosphere-ionosphere interactions to derive electric field data near Earth’s boundary. A simplified ionospheric model for the F-layer was implemented to compute the O+ and electron densities. Subsequently, the spacecraft drag coefficient and the plasma wake size were determined for varying solar wind speed inputs corresponding to quiet, intense, and extreme space weather conditions. In the case of extreme conditions with a solar wind speed of 1000 km/s, the results indicated that the excess satellite drag coefficient and plasma wake size increased to around ten percent.

Digital Application of Folk Culture Elements in City Branding Visual Identity Shaping

Abstract In this paper, for the visual image shaping of folk culture elements in city branding, the intelligent extraction technology of digital image with image pre-processing, color space conversion, and color extraction as the three core modules is proposed. In terms of image preprocessing, spatial domain filtering is used as a means to explore the two main algorithms of linear filtering and nonlinear filtering, and in the color space conversion module, the RGB color space is converted to the CIE Lab color space with the help of CIE XYZ color. The color map extraction module uses Python language to construct the color collocation relationship network using K-means and color adjacency matrix visualization. Finally, the quantitative experiment on the effect of the city brand visual image that integrates the elements of folk culture is carried out, and 65 city brand visual image samples are quantitatively analyzed, while the AMOS model is applied to test the city brand visual image.

Towards addressing aggregation deviation for model training in resource-scarce edge environment

Federated Learning (FL) can train models in an edge environment without sending raw data. However, the performance is still constrained by data heterogeneity. To address the problems of data heterogeneity and resource scarcity in edge devices, we propose Federal Learning via Dynamic Aggregation (FedDA), which eliminates the influence of data heterogeneity and improves model accuracy. FedDA updates the impact of individual local models on the global model in real-time at different stages. It adjusts the local epoch in each round to prevent the device from dropping out while obtaining a more accurate local model. The core module is the model impact factor (MIF) that inscribes the aggregation weights to solve the impact of fixed weights on the aggregation model with improper extraction of local information. We conducted several experiments to evaluate the convergence speed using different algorithms on the MINIST. FedDA consistently outperforms the other six SOTA algorithms on MNIST, Cifar10, and Cifar100 datasets. In significant data heterogeneity, FedDA improves accuracy by up to 6% over the different algorithms and at least about 3%, especially in resource-scarce environments. To reach the specified accuracy, FedDA is 3 times faster than SCAFFOLD and at least 50% faster than other algorithms.

A versatile micromodel technology to explore biofilm development in porous media flows.

Bacterial biofilms that grow in porous media are critical to ecosystem processes and applications ranging from soil bioremediation to bioreactors for treating wastewater or producing value-added products. However, understanding and engineering the complex phenomena that drive the development of biofilms in such systems remains a challenge. Here we present a novel micromodel technology to explore bacterial biofilm development in porous media flows. The technology consists of a set of modules that can be combined as required for any given experiment and conveniently tuned for specific requirements. The core module is a 3D-printed micromodel where biofilm is grown into a perfusable porous substrate. High-precision additive manufacturing, in particular stereolithography, is used to fabricate porous scaffolds with precisely controlled architectures integrating flow channels with diameters down to several hundreds of micrometers. The system is instrumented with: ultraviolet-C light-emitting diodes; on-line measurements of oxygen consumption and pressure drop across the porous medium; camera and spectrophotometric cells for the detection of biofilm detachment events at the outlet. We demonstrate how this technology can be used to study the development of Pseudomonas aeruginosa biofilm for several days within a network of flow channels. We find complex dynamics whereby oxygen consumption reaches a steady-state but not the pressure drop, which instead features a permanent regime with large fluctuations. We further use X-ray computed microtomography to image the spatial distribution of biofilms and computational fluid dynamics to link biofilm development with local flow properties. By combining the advantages of additive manufacturing for the creation of reproducible 3D porous microarchitectures with the flow control and instrumentation accuracy of microfluidics, our system provides a platform to study the dynamics of biofilm development in 3D porous media and to rapidly test new concepts in process engineering.

Spectral Super-Resolution via Deep Low-Rank Tensor Representation.

Spectral super-resolution has attracted the attention of more researchers for obtaining hyperspectral images (HSIs) in a simpler and cheaper way. Although many convolutional neural network (CNN)-based approaches have yielded impressive results, most of them ignore the low-rank prior of HSIs resulting in huge computational and storage costs. In addition, the ability of CNN-based methods to capture the correlation of global information is limited by the receptive field. To surmount the problem, we design a novel low-rank tensor reconstruction network (LTRN) for spectral super-resolution. Specifically, we treat the features of HSIs as 3-D tensors with low-rank properties due to their spectral similarity and spatial sparsity. Then, we combine canonical-polyadic (CP) decomposition with neural networks to design an adaptive low-rank prior learning (ALPL) module that enables feature learning in a 1-D space. In this module, there are two core modules: the adaptive vector learning (AVL) module and the multidimensionwise multihead self-attention (MMSA) module. The AVL module is designed to compress an HSI into a 1-D space by using a vector to represent its information. The MMSA module is introduced to improve the ability to capture the long-range dependencies in the row, column, and spectral dimensions, respectively. Finally, our LTRN, mainly cascaded by several ALPL modules and feedforward networks (FFNs), achieves high-quality spectral super-resolution with fewer parameters. To test the effect of our method, we conduct experiments on two datasets: the CAVE dataset and the Harvard dataset. Experimental results show that our LTRN not only is as effective as state-of-the-art methods but also has fewer parameters. The code is available at https://github.com/renweidian/LTRN.

Qualidade de vida relacionada à saúde de uma criança com Trissomia 21 e Leucemia Mieloide Aguda: relato de caso

ABSTRACT Trisomy 21 (Down syndrome) is a genetic condition arising from the presence of three chromosomes at position 21. Individuals diagnosed with this syndrome have a greater likelihood of developing systemic problems, as well as cognitive, developmental, and psychiatric disorders. This paper reports a case of a child with Down syndrome under treatment for acute myeloid leukemia, emphasizing its impact on quality of life. The mother’ perception on health-related quality of life (HRQoL) was measuredby the Pediatric Quality of Life Inventory (PedsQL), Generic Core and Cancer Module Scales. The mother’s perception shown that physical and functional functioning indicated a higher negative impact on the child’s life. The lowest scores were in the domains nausea and anxiety about the procedure. Intraoral examination revealed an incomplete primary dentition with ectopic eruption. The child has had a multidisciplinary health care approach (pediatrician, hematologist, gastroenterologist, cardiologist, physiotherapist and dentist). Therefore, it is essential that health professionals, including dentists, to incorporateinto their practice the knowledge on the difficulties and emotional aspects of this population, providinga patient-centered humanized care.

Protocol for the Swiss COhort of Healthcare Professionals and Informal CAregivers (SCOHPICA): Professional trajectories, intention to stay in or leave the job and well-being of healthcare professionals.

Healthcare professionals' shortage, low job satisfaction, high levels of burnout, and excessive staff turnover are some of the challenges health systems face worldwide. In Switzerland, healthcare stakeholders have called to address the health workforce crisis and have pointed out the scarcity of data on the conditions of healthcare professionals (HCPs). Hence, the Swiss Cohort of Healthcare Professionals and Informal Caregivers (SCOHPICA) was developed to study the career trajectories, well-being, intention to stay in or leave the position/profession/health sector, and their determinants, of HCPs and informal caregivers, respectively. This paper describes the protocol for the HCPs cohort of SCOHPICA and discusses its implications. SCOHPICA is a prospective open cohort using an explanatory sequential mixed methods design. All types of HCPs working directly with patients and practicing in Switzerland are eligible, irrespective of their healthcare setting and employment status. Baseline and annual follow-up electronic surveys will take place once a year, featuring both core questions and modules developed according to information needs. While outcome variables are HCPs' trajectories, well-being, intention to stay in or leave the position/profession/health sector, independent variables include organizational, psychosocial, and psychological determinants, as well as occupational (professional) and sociodemographic factors. The qualitative phase will be organized every two years, inviting participants who agreed to take part in this phase. The findings from quantitative analyses, along with the issues raised by healthcare stakeholders in the field, will guide the topics investigated in the qualitative phase. Using innovative methodologies, SCOHPICA will gather nationwide and longitudinal data on HCPs practicing in Switzerland. These data could have numerous implications: promoting the development of research related to HCPs' well-being and retention intentions; supporting the development of policies to improve working conditions and career prospects; contributing to the evolution of training curricula for future or current healthcare professionals; aiding in the development of health systems capable of delivering quality care; and finally, providing the general public and stakeholders with free and open access to the study results through an online dashboard.

Defining dignity in higher education as an alternative to requiring 'Trigger Warnings'.

This article examines trigger warnings, particularly the call for trigger warnings on university campuses, and from a Levinasian and Kantian ethical perspective, and addresses the question: When, if ever, are trigger warnings helpful to student's learning? The nursing curriculum is developed with key stakeholders and regulatory bodies to ensure graduate nurses are competent to deliver a high standard of care to patients and clients. Practical teaching practice and published research has uncovered an increasing use of 'Trigger Warnings' before a topic is discussed, or used as warnings on core module texts. It is appreciated that some students have personal experience of psychological or physical trauma. However, apart from identifying these students through Mitigating Circumstances committees, or when the student feels confident to share this information with a personal tutor, this information remains strictly confidential. There is the potential for covert skills such as critical analysis and skilful discussion not being attained by the student. With the assistance of Kants moral theory, an argument will develop that the insidious use of Trigger warnings and the embargo of recommended reading, requires critical discussion with the public. This would involve the rationale and pedagogical justification for the use of texts, and the necessity within nursing education to address challenging clinical topics. To support students with PTSD this may involve the research discussed on personal educational needs analysis.

Artificial Intelligence-Assisted Training and Skill Enhancement Strategies for Athletes

Abstract The continuous progress of science and technology promotes the continuous development of artificial intelligence, and digital intelligence empowerment has gradually become a focus topic of widespread social concern. Considering that traditional sports training is difficult to meet the higher requirements put forward by athletes on training methods, training load, and training equipment, this study combines artificial intelligence technology to explore the path of skill enhancement. In this paper, based on the skeletal key points for athlete training action recognition, the improved DHRNet network is optimized for the detection of skeletal key points and then for the complexity of the athlete’s training state, the occlusion of the skeletal points and the repair of motion information are carried out. On this basis, we construct an artificial intelligence-assisted training system composed of a video image input function system, action recognition function system, and behavior analysis function system. After completing the design, the core module of the system, the motion recognition system, was tested. The average recognition success rates of basic and sport movements were between 90.8~95.6% and 84.5~96.5%, respectively. The overall recognition correctness rate remained above 90% in the actual application test, which can meet the training needs of athletes. The average score for all abilities of the athletes using the AI-supported system is 85.242, which is 6.21 higher than that of the traditional training method. The AI-assisted system is of great help to athletes’ training, and it also enables the early realization of intelligent training.

Dynamic Interaction Dilation for Interactive Human Parsing

Interactive segmentation pursues generating high-quality pixel-level predictions with a few user-provided clicks, which is gaining attention for its convenience in segmentation data annotation. Users are allowed to iteratively refine the prediction by adding clicks until the result is satisfactory. Existing interactive methods usually transform the clicks into a set of localization maps by Euclidian distance computation or RGB texture extraction to guide the segmentation, which makes the click transformation a core module in interactive segmentation networks. However, when adopted in human images where large poses, occlusions, and bad illuminations are prevailing, prior transformation methods tend to cause uncorrectable overlapping across localization maps, <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">i.e.</i> , one click corresponds to multiple transformed values at the same position in different localization map channels, which are difficult to form a good match among human parts and limit the interaction efficiency. Furthermore, the inappropriately transformed information is hard to be refined with the static transformation manner, <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">i.e.</i> , based on the fixed formulas / RGB textures, which is out of tune with the dynamically refined interaction process. Hence, we design a dynamic transformation scheme for interactive human parsing (IHP) named Dynamic Interaction Dilation Net ( <bold xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">DID-Net</b> ), which serves as an initial attempt to break the limitations of static transformation while capturing long-range dependencies of clicks within each human part. Specifically, we construct a Dynamic Dilation Module ( <bold xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">DD-Module</b> ) to dilate clicks radially in several directions assisted by human body edge detection. The continually refined edges guide to improve the dilation quality in each interaction iteration, thereby better fitting user intention. Furthermore, we propose an Adaptive Interaction Excitation Block ( <bold xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">AIE-Block</b> ) to exploit potential semantic clues buried in the dilated clicks and emphasize semantic expression for each human part by feature recalibration. Our DID-Net achieves state-of-the-art performance on 3 public human parsing benchmarks.

A General Spatial-Frequency Learning Framework for Multimodal Image Fusion.

multimodal image fusion involves tasks like pan-sharpening and depth super-resolution. Both tasks aim to generate high-resolution target images by fusing the complementary information from the texture-rich guidance and low-resolution target counterparts. They are inborn with reconstructing high-frequency information. Despite their inherent frequency domain connection, most existing methods only operate solely in the spatial domain and rarely explore the solutions in the frequency domain. This study addresses this limitation by proposing solutions in both the spatial and frequency domains. We devise a Spatial-Frequency Information Integration Network, abbreviated as SFINet for this purpose. The SFINet includes a core module tailored for image fusion. This module consists of three key components: a spatial-domain information branch, a frequency-domain information branch, and a dual-domain interaction. The spatial-domain information branch employs the spatial convolution-equipped invertible neural operators to integrate local information from different modalities in the spatial domain. Meanwhile, the frequency-domain information branch adopts a modality-aware deep Fourier transformation to capture the image-wide receptive field for exploring global contextual information. In addition, the dual-domain interaction facilitates information flow and the learning of complementary representations. We further present an improved version of SFINet, SFINet++, that enhances the representation of spatial information by replacing the basic convolution unit in the original spatial domain branch with the information-lossless invertible neural operator. We conduct extensive experiments to validate the effectiveness of the proposed networks and demonstrate their outstanding performance against state-of-the-art methods in two representative multimodal image fusion tasks: pan-sharpening and depth super-resolution. The source code is publicly available at https://github.com/manman1995/Awaresome-pansharpening.

Future sustainable space exploration: Space conditions adaptation and self-sustaining system

Creating regenerative systems onboard space vehicles is the next technological breakthrough to reduce cost, efficiency, and sustainability of space exploration. Near future projects include NASA’s plan to set up moon base as soon as 2028 in one of the Artemis missions. Upon architectural and structural maturation, focus is shifted to maintaining presence. This includes minimising astrological impacts on the crew and building self-sustaining systems that would be the key to a new era of ambitious missions. ALS and HLS will provide transportation for the missions from the Earth to destinations, with impacts on the crew during journey discussed. Ways of minimising effects are vital to missions in the long run. Process of activation and operating core modules is portrayed on blueprints. Details of development and current technological advances that could be utilised may offer potential on the system, new advancements by the time of liftoff is expected for taking pioneering space missions to a new level.

Construction of personalized genomics model and clinical application in precision medicine

Abstract Sequencing technology continues to evolve, and pharmacogenomics is increasingly pivotal in the pursuit of personalized medicine. This study delineates the personalized genomics model into two core modules for framework construction: data preprocessing and prediction. Within the data preprocessing module, a denoising submodule and a genomic feature distribution alignment module are dedicated to processing the genomic features associated with diseases and assimilating them into the spatial feature distribution of the model. The prediction module employs a fully connected neural network alongside a graph convolutional neural network to forecast drug dosages based on the disease's genomic features. We propose clinical applications of personalized genomics models across three distinct pathways. To assess the practical impact of these models, experiments were conducted focusing on their clinical application. After 14 days of medication administration, the International Normalized Ratio (INR) value for the drug model group reached 2.67, surpassing that of the conventional treatment group, with nearly 60% of participants achieving the target range. Furthermore, in evaluating the safety and efficacy of the clinical application of the genomics model, the incidence of serious bleeding events in the drug model group was recorded at 16.67%, which is below the standard allowable probability of 23.81%. Therefore, the personalized genomics model introduced in this study meets the established safety and efficacy standards.

Towards Unsupervised Referring Expression Comprehension with Visual Semantic Parsing

Referring Expression Comprehension (REC) is a task that involves grounding a specific object in an image based on a given referring query in the form of bounding boxes. Existing fully-supervised or weakly-supervised REC methods rely on either manually annotated regional coordinates or query texts. In this paper, we propose an unsupervised training paradigm for the REC task that does not require any manual annotated data. Specifically, we introduce a Visual-Semantic-Parsing-based Unsupervised Referring Expression Comprehension framework (VUREC), which leverages a Visual Semantic Parser (VSP) as its core module to recognize the rich semantic information in images and construct pseudo-region-query pairs as the training supervision for REC. The VSP comprises a Scene Graph Parser (SGP) and a Visual Concept Detector (VCD) that can detect the locations, categories, attributes of objects, and visual relationships among them in images. Furthermore, we present a Referring Expression Reasoning (RER) model that utilizes a Multi-Modal Cascade Attention Decoder (MCAD) for fine-grained multi-modality fusion and regresses the four coordinates of the referential object directly. The experimental results on three benchmark datasets of Refcoco, Refcoco+ and Refcocog demonstrate the effectiveness of our proposed method.

Skills-based online self-help program for carers of people with an eating disorder (SupportED): A case series evaluation.

To investigate the acceptability, feasibility, and preliminary effectiveness of "SupportED," a novel, online, self-paced skills-based program for carers/support people of those with lived and/or living experience of an eating disorder (ED). Carers of a loved one with an ED were invited to undertake a five-module skills-based (one core module, four optional modules) online program at their own pace. Modules cover general information about EDs, communication, practical skills, treatment engagement, and recovery. Participants completed demographics and were assessed using EDSIS, CASK, and DASS-21 for caregiver burden, skills, and distress, respectively, at program completion and 3-month follow-up. Feedback on the program and modules was also collected. Most carers (65%; 82/126) completed the core module of the program, which is on par with other online learning programs. Nearly half of program completers responded to the 3-month follow-up (n = 39). On average, carers rated the program as relevant and useful and found the information validating and easy to understand, albeit a few carers found the program time-consuming and technically challenging. Intent-to-treat (ITT) linear mixed models demonstrated the preliminary effectiveness of the program for improving carer burden and skills, although the change in carer distress was non-significant. Results of this case series evaluation suggest SupportED is an acceptable and feasible program for carers and may improve burden and support skill development. Although the online, self-paced format was highly rated, small improvements to improve acceptability were suggested. Reasons for a lack of impact on carer distress are discussed. This article describes a case series evaluation of "SupportED," an online skills-based support program for families, carers, and supporters of people with lived experience of an eating disorder. This is the first online carer support program that is designed to be inclusive of all carer types (e.g., partners, siblings) for all eating disorder presentations. Findings suggest the program's suitability and helpfulness for reducing carers' load and bolstering skills and knowledge.

Design of a Novel Magnetic Induction Switch with a Permalloy Film and a Trans-Impedance Amplifier Circuit

At present, magnetic induction switches are widely used in industrial automation control and biological sensing systems. A core module composed of a magnetic sensing device and a signal conditioning circuit is designed and analyzed in this paper. Utilizing a permalloy film with the anisotropic magneto-resistance (AMR) effect, the novel magnetic induction switch shows its ability to correctly detect the direction of magnetic fields. Furthermore, an interfacial circuit based on a trans-impedance amplifier (TIA) is designed to measure and regulate the output signal of the sensing device. Accurate simulation results show the gain of the TIA reaches up to 51.36 dB with a bandwidth of 1.3 GHz and a power consumption of 3.65 mW. The outstanding performance of the proposed module demonstrates the possibility of solving the problems induced by high input impedance, high frequency, and parasitic effects in magnetic induction switches.

Multimodal multiscale dynamic graph convolution networks for stock price prediction

Predicting directional future stock price movements is very challenging due to the complex, stochastic, and evolving nature of the financial markets. Existing literature either neglects other timely and granular alternative data, such as media text data, or fails to extract and distill predictive multimodal features from the data. Moreover, the time-varying cross-sectional relations beyond sequential dependencies of stock prices are informative for forecasting price fluctuations, for which the modelling flexibility, however, is not adequate in most of the previous studies. In this paper, we propose a novel Multiscale Multimodal Dynamic Graph Convolution Network (Melody-GCN) to address these issues in stock price prediction. It contains three core modules: (1) multimodal fusing-diffusing blocks that effectively integrate and align the numerical and textual features; (2) a multiscale architecture that extracts and refines temporal features via a fine-to-coarse descending path and a coarse-to-fine ascending path progressively; and (3) dynamic spatio-temporal graph convolutional layers that learn the complex and evolving stock relations not only in between industries and individual companies but also across time horizons. Extensive experimental results and trading simulations on two real-world datasets demonstrate the superior performance of our proposed approach beyond other state-of-the-art models.

Arabidopsis BBX14 is involved in high light acclimation and seedling development.

The development of photosynthetically competent seedlings requires both light and retrograde biogenic signaling pathways. The transcription factor GLK1 functions at the interface between these pathways and receives input from the biogenic signal integrator GUN1. BBX14 was previously identified, together with GLK1, in a core module that mediates the response to high light (HL) levels and biogenic signals, which was studied by using inhibitors of chloroplast development. Our chromatin immunoprecipitation-Seq experiments revealed that BBX14 is a direct target of GLK1, and RNA-Seq analysis suggests that BBX14 may function as a regulator of the circadian clock. In addition, BBX14 plays a role in chlorophyll biosynthesis during early onset of light. Knockout of BBX14 results in a long hypocotyl phenotype dependent on a retrograde signal. Furthermore, the expression of BBX14 and BBX15 during biogenic signaling requires GUN1. Investigation of the role of BBX14 and BBX15 in GUN-type biogenic (gun) signaling showed that the overexpression of BBX14 or BBX15 caused de-repression of CA1 mRNA levels, when seedlings were grown on norflurazon. Notably, transcripts of the LHCB1.2 marker are not de-repressed. Furthermore, BBX14 is required to acclimate plants to HL stress. We propose that BBX14 is an integrator of biogenic signals and that BBX14 is a nuclear target of retrograde signals downstream of the GUN1/GLK1 module. However, we do not classify BBX14 or BBX15 overexpressors as gun mutants based on a critical evaluation of our results and those reported in the literature. Finally, we discuss a classification system necessary for the declaration of new gun mutants.