Task Success Research Articles

Neighbors selective Graph Convolutional Network for homophily and heterophily

Graph Convolutional Networks (GCNs) gain remarkable success in graph-related tasks under homophily graph assumption—most connected nodes have the same label. However, this assumption is fragile since heterophily is common in real-world networks, where most linked nodes have different labels. Some methods address heterophily using graph-level weighted combinations of multi-hop neighbor representations. However, they introduce noise and irrelevant information from two aspects: (1) higher-order neighborhoods may include more neighbors with different labels than those with the same label; (2) graph-level weighted combination fails to capture distinct node properties. To address these issues, firstly, a Neighborhood Distribution-induced Similarity (NDS) measure is developed to identify potential similar neighbors for nodes. Secondly, a node-level fusion mechanism Selective-Neighbors Gated Unit (SNGU) is designed to adaptively aggregate potential neighbors, first-hop neighbors, and the node itself by learning distinct weights for nodes. By combining the above two designs, this work proposes a novel Neighbors Selective Graph Convolutional Network (NSGCN), which allows nodes to selectively receive relevant neighbor information for better node representations, effectively modeling homophily and heterophily. Experiments on 10 widely used real-world datasets with varying properties demonstrate the superiority of the NSGCN, surpassing a strong baseline with an absolute increase of 9.83% in classification accuracy. The code is available at https://github.com/GGA23/NSGCN.

Cost-aware task offloading in vehicular edge computing: A Stackelberg game approach

With the popularity of vehicular communication systems and mobile edge vehicle networking, intelligent transportation applications arise in Internet of Vehicles (IoVs), which are latency-sensitive, computation-intensive, and requiring sufficient computing and communication resources. To satisfy the requirements of these applications, computation offloading emerges as a new paradigm to utilize idle resources on vehicles to cooperatively complete tasks. However, there exist several obstacles for realizing successful task offloading among vehicles. For one thing, extra cost such as communication overhead and energy consumption occurs when a task is offloaded on a service vehicle, it is unlikely to expect the service vehicle will contribute its resources without any reward. For another, since there are many vehicles around, both user vehicles and service vehicles are trying to strike a balance between cost and profit, through matching the perfect service/user vehicles and settled with optimal offloading plan that is beneficial to all parties. To solve these issues, this work focuses on the design of effective incentive mechanisms to stimulate vehicles with idle resources to actively participate in the offloading process. A fuzzy logic-based dynamic pricing strategy is proposed to accurately evaluate the cost of a vehicle for processing the task, which provides insightful guidance for finding the optimal offloading decision. Meanwhile, the competitive and cooperation relations among vehicles are thoroughly investigated and modeled as a two-stage Stackelberg game. Particularly, this work emphasizes the social attributes of vehicles and their effect on the offloading decision making process, multiple key properties such as the willingness of UV to undertake the task locally, the reputation of UV and the satisfaction of SV for the allocated task proportion, are carefully integrated in the design of the optimization problem. A distributed algorithm with applicable complexity is proposed to solve the problem and to find the optimal task offloading strategy. Extensive simulations are conducted on real-world scenarios and results show that the proposed mechanism achieves significant performance advantages in terms of vehicles' utilities, cost, completion delay under varied network and channel environment, which justifies the effectiveness and efficiency of this work.

Multiobjective Evolutionary Generative Adversarial Network Compression for Image Translation

Generative Adversarial Networks (GANs) have achieved remarkable success in image translation tasks. However, its prohibitive computational overhead has been a major hurdle for deployment on resource constrained platforms. Due to the training instability and the complicated network architecture, existing model acceleration techniques cannot appropriately handle the GAN compression problem. To cope with these difficulties, we propose a multi-objective evolutionary algorithm to compress GAN models in the context of image translation tasks, which is termed as MEGC. Particularly, the conflict between the computational cost of the GAN model and the quality of generated image is explicitly modeled as a two-objective optimization problem, and the evolved Pareto set is utilized to guide the sampling process during supernet training, which can in turn divert the focus of the supernet training to well-performing compact subnets. Besides, an evaluation-free strategy is introduced to facilitate exploration in the search space while incurring no extra computational cost. Based on the above design, the proposed MEGC eliminates the requirement of subnet searching in the post-processing procedure. Experiments on image translation tasks under paired and unparied settings demonstrate the effectiveness of the proposed MEGC on reducing the computational cost of GANs while improving the quality of generated images compared to those of the full models.

Learning Graph Representations Through Learning and Propagating Edge Features.

Graph convolutional networks have achieved considerable success in various graph domain tasks. Recently, numerous types of graph convolutional networks have been developed. A typical rule for learning a node's feature in these graph convolutional networks is to aggregate node features from the node's local neighborhood. However, in these models, the interrelation information between adjacent nodes is not well-considered. This information could be helpful to learn improved node embeddings. In this article, we present a graph representation learning framework that generates node embeddings through learning and propagating edge features. Instead of aggregating node features from a local neighborhood, we learn a feature for each edge and update a node's representation by aggregating local edge features. The edge feature is learned from the concatenation of the edge's starting node feature, the input edge feature, and the edge's end node feature. Unlike node feature propagation-based graph networks, our model propagates different features from a node to its neighbors. In addition, we learn an attention vector for each edge in aggregation, enabling the model to focus on important information in each feature dimension. By learning and aggregating edge features, the interrelation between a node and its neighboring nodes is integrated in the aggregated feature, which helps learn improved node embeddings in graph representation learning. Our model is evaluated on graph classification, node classification, graph regression, and multitask binary graph classification on eight popular datasets. The experimental results demonstrate that our model achieves improved performance compared with a wide variety of baseline models.

Alzheimer’s Disease Diagnosis via SVM-Driven Classification of Structural MRI Data

Abstract: In this study we introduce an automated system designed to classify Alzheimers disease (AD) using MRI brain scans. Our technique employs support vector machine (SVM) classification to effectively differentiate between AD patients and older adults without the disease by analyzing brain MRI images. The process includes dividing 2D T1 weighted MR images into regions of interest and extracting gray matter characteristics, from each region. By leveraging SVMs reliability and proven success in classification tasks along with the principle of minimizing risks our research showcases performance in generalizing results. This SVM based approach shows promise in distinguishing AD affected brains from ones presenting a valuable opportunity, for enhancing medical analysis and interpretation of MRI findings.

SKATE : Successive Rank-based Task Assignment for Proactive Online Planning

SKATE : Successive Rank-based Task Assignment for Proactive Online Planning

Success in the Natural Detection Task is influenced by only a few factors generally believed to affect dogs’ olfactory performance

Research into dogs' olfactory ability is growing rapidly. However, generalising based on scientific results is challenging, because research has been typically conducted on a few specially trained subjects of a few breeds tested in different environmental conditions. We investigated the effects of temperature and humidity (outdoors), age, test location, sex, neutering status, and repeated testing (outdoors and indoors) on the olfactory performance of untrained family dogs (N = 411) of various breeds. We employed the Natural Detection Task with three difficulty levels, from which we derived two performance metrics: Top Level and Success Score. Temperature (0–25 °C) and humidity (18–90%) did not affect olfactory performance. Young adult dogs surpassed other age groups in reaching the Top Level. Sex and neutering status showed no discernible influence on Top Level and Success Score. Dogs performed better in both metrics when tested indoors compared to outdoors. In the test–retest procedure no significant learning effect was observed. We confirmed on untrained companion dogs that olfactory performance declines with age and rejected some factors that have been previously hypothesised to significantly affect dogs’ olfactory success. The influence of the testing environment was notable, emphasising the need to consider various factors in understanding dogs' olfactory capabilities.

Indoor Localization and Path Planning of Mobile Disinfection and Epidemic Prevention Robot

Introduction: In response to the challenge of autonomous localization and navigation for mobile disinfection robots in unknown environments during the COVID-19 pandemic based on the existing patent technology of mobile robots, a localization and navigation approach combining the Cartographer-Simultaneous Localization and Mapping (SLAM) algorithm, an improved A* algorithm, and the Dynamic Window Approach (DWA) has been proposed, enabling effective disinfection and prevention work. Method: First, a comprehensive analysis of two laser SLAM algorithm frameworks, Gmapping- SLAM and Cartographer-SLAM, was conducted, which determined the Cartographer algorithm as the optimal SLAM algorithm for the mobile disinfection robot due to its superior performance. Second, an improved A* algorithm was developed based on its principles and integrated with the DWA algorithm for optimal path planning. Result: Simulation and prototype experiments demonstrated that the proposed approach effectively solved the autonomous localization and navigation challenges faced by mobile disinfection robots in unknown environments, enabling successful disinfection and prevention tasks. Conclusion: The proposed approach has the potential to lead to new patent applications in this field.

The essence of motivation in learning English as a foreign or second language

This research paper investigates recent studies on the role of motivation in second or foreign language acquisition. Motivation is identified as a critical determinant of success or failure in complex tasks. The paper offers academic definitions of motivation, categorizes different types of motivation, and revisits prior research on its significance in language learning. It also outlines the essential requirements and key aspects of motivation. The literature review underscores the pivotal role of motivation in learning English as a second or foreign language, revealing that motivation significantly enhances language learning outcomes. The findings emphasize that fostering motivation is essential for achieving effective and successful language acquisition. This comprehensive examination provides valuable insights into how motivation can be leveraged to improve language learning processes and outcomes. Keywords: English as a foreign language; impact; motivation; second language.

An Imaging-Compatible Oral Retractor System for Transoral Robotic Surgery.

This study aimed to develop and validate a Computed Tomography (CT)/Magnetic Resonance Imaging(MRI)-compatible polymer oral retractor system to enable intraoperative image guidance for transoral robotic surgery (TORS). The retractor was designed based on standard-of-care metallic retractors and 3D (three-dimensional) printed with carbon fiber composite and nylon. The system was comprehensively evaluated in bench-top and cadaveric experiments in terms of its ability to enable intraoperative CT/MR images during TORS, functionality including surgical exposure and working volume, usability, compatibility with da Vinci surgical systems, feasibility for disinfection or sterilization, and robustness over an extended period of time. The polymer retractor system enabled the acquisition of high-resolution and artifact-free intraoperative CT/MR images during TORS. With an inter-incisive distance of 42.55mm and a working volume of 200.09cm3, it provided surgical exposure comparable to standard-of-care metallic retractors. The system proved intuitive and compatible with da Vinci S, Xi, and Single Port systems, enabling successful mock surgical tasks performed by surgeons and residents. The retractor components could be effectively disinfected or sterilized for clinical use without significant compromise in material strength, with STERRAD considered the optimal method. Throughout a 2h mock procedure, the retractor system showed minimal displacements (<1.5mm) due to surrounding tissue deformation, with insignificant device deformation. The 3D-printed polymer retractor system successfully enabled artifact-free intraoperative CT/MR imaging in TORS for the first time and demonstrated feasibility for clinical use. This breakthrough opens the door to surgical navigation with intraoperative image guidance in TORS, offering the potential to significantly improve surgical outcomes and patients' quality of life.

MDCNet: Long-term time series forecasting with mode decomposition and 2D convolution

Long-term time series forecasting is widely used in various real-world applications, such as weather, traffic, energy, healthcare, etc. Recently, time series decomposition techniques have been adopted in many mainstream forecasting models, such as the prevalent Transformer-based models, to help capture sophisticated temporal patterns and achieve success in several benchmark tasks. However, conventional decomposition algorithms are often based on simple operations or limited to specific fields, and therefore are not effective and applicable enough, especially for complex time series. In this paper, we propose Mode Decomposition and 2D Convolutional Network (MDCNet), a structure-simple yet effective forecasting architecture based on a more effective decomposition method and a multi-frequency time series feature extraction network with multi-scale 2D convolution. Specifically, we first introduce a Variational Mode Decomposition Block to discover intricate time patterns, which decompose time series into trend components and stationary modal components at different main frequencies. Then, we design a Trend Prediction Block and an Intrinsic Mode Functions Prediction Block to capture global correlation and hidden dependencies within different main frequencies, respectively. Furthermore, a Frequency Enhancement Module is designed to further reduce the impact of noise in long-term time series. Experiments on eight benchmark datasets show that MDCNet significantly reduces the error of the previous state-of-the-art method by 15.1% and 11.5% for multivariate and univariate time series, respectively.

Structure enhanced prototypical alignment for unsupervised cross-domain node classification

Graph Neural Networks (GNNs) have demonstrated remarkable success in graph node classification task. However, their performance heavily relies on the availability of high-quality labeled data, which can be time-consuming and labor-intensive to acquire for graph-structured data. Therefore, the task of transferring knowledge from a label-rich graph (source domain) to a completely unlabeled graph (target domain) becomes crucial. In this paper, we propose a novel unsupervised graph domain adaptation framework called Structure Enhanced Prototypical Alignment (SEPA), which aims to learn domain-invariant representations on non-IID (non-independent and identically distributed) data. Specifically, SEPA captures class-wise semantics by constructing a prototype-based graph and introduces an explicit domain discrepancy metric to align the source and target domains. The proposed SEPA framework is optimized in an end-to-end manner, which could be incorporated into various GNN architectures. Experimental results on several real-world datasets demonstrate that our proposed framework outperforms recent state-of-the-art baselines with different gains.

External characteristics of cardio-strength load based on local exercises and heart rate response during its performance

In modern health-enhancing physical culture the use of complex training sessions with successive tasks solving is mainly provided and almost no attention to complex training sessions with parallel task solving. In this regard, it is important to find out the possibility of building training sessions that would allow influencing the cardio respiratory system through the use of strength exercises. The purpose of the study is to explore the characteristics of the load and heart rate changing in complex cardio-strength training sessions with parallel task solving. Experimental total training sessions included a warm-up (10-12 min) and 6 local strength exercises divided into three pairs (biceps - triceps, deltoid - latissimus dorsi, rhomboid - pectoral muscles) performed as circuit training. The duration of cardio- strength sessions – 30 min., intensity – 12 RM, repetition – 4, rest – changing exercises. A 62 years old man with high physical fitness for this age took place in a research. The heart rate study was carried out by the POLAR cardio-tracker. The sound recording of the exercise commentary was simultaneously activated with the heart rate sensor. This made it possible to study the duration of exercises, rests and heart rate during each exercise. Research has shown it is possible to ensure a cardiovascular reaction appropriate for cardio training by using strength exercises. But it requires certain resistance, repetition and rest between exercises at the training session. It is advisable to call such loads as cardio-strength loads. A person is able to provide the necessary duration of exercises and rests in such training sessions without a clock or sound leader. The current heart rate during certain cardio-strength exercise is determined by the complex, ambiguous interaction of several factors. For loads of 1/3 of 12 RM (4 repetitions per set) the most impotent is the initial position. In the initial position the most influential is the body position while performing exercises (lying, sitting, bending over and standing); the activity of the synergist muscles is also important here. In addition, the heart rate reaction depends on a resistance, the preliminary exercise and the activity during the changing exercises.

The effect of magnesium sulfate on memory and anxiety-like behavior in a rat model: an investigation of its neuronal molecular mechanisms

ABSTRACT Background Anxiety is an adaptive response to potentially threatening conditions. Excessive and uncontrolled anxiety responses become nonadaptive and cause anxiety disorders. To better understand the anxiety-modulating effects of Mg sulfate, behavioral test batteries in the assessment of anxiety and learning and memory functions were performed simultaneously over a time period. This study also examines the effects of Mg sulfate compared to diazepam, an anxiolytic drug with amnestic effects on anxiety-like behavior, as well as possible oxidative-nitrosative stress and hippocampal changes in male rats exposed to predator odor. Methods Young adult Sprague-Dawley male rats were used. The rats were assessed using a comprehensive neurobehavioral test battery consisting of novel object recognition, open field, and successive alleys tasks. Anxiety was induced by cat odor, and diazepam and Mg were used as study drugs. Of the frontal cortex and hippocampus, the state of total oxidant and antioxidant and NO levels and histological examination of hippocampal CA1, CA2, CA3, and DG regions were performed. Results Diazepam- and Mg-treated rats showed an improvement in anxiety-related behavior to predator odors. Furthermore, Mg treatment alleviated some of the increasing oxidative stress in the frontal cortex and hippocampus of rats, while diazepam treatment in particular enhanced hippocampal oxidant and antioxidant activity. In addition, brain NO increase induced by animal odor exposure or diazepam treatment was ameliorated by Mg administration. Conclusions Overall, our work suggests that Mg had a partial anxiolytic effect on anxiety-like behaviors, although not as much as diazepam, and this effect varied depending on the dose. Mg treatment might counteract increased oxidative stress and elevated NO levels in the brain.

CrisisTransformers: Pre-trained language models and sentence encoders for crisis-related social media texts

Social media platforms play an essential role in crisis communication, but analysing crisis-related social media texts is challenging due to their informal nature. Transformer-based pre-trained models like BERT and RoBERTa have shown success in various NLP tasks, but they are not tailored for crisis-related texts. Furthermore, general-purpose sentence encoders are used to generate sentence embeddings, regardless of the textual complexities in crisis-related texts. Advances in applications like text classification, semantic search, and clustering contribute to the effective processing of crisis-related texts, which is essential for emergency responders to gain a comprehensive view of a crisis event, whether historical or real-time. To address these gaps in crisis informatics literature, this study introduces CrisisTransformers, an ensemble of pre-trained language models and sentence encoders trained on an extensive corpus of over 15 billion word tokens from tweets associated with more than 30 crisis events, including disease outbreaks, natural disasters, conflicts, and other critical incidents. We evaluate existing models and CrisisTransformers on 18 crisis-specific public datasets. Our pre-trained models outperform strong baselines across all datasets in classification tasks, and our best-performing sentence encoder improves the state-of-the-art by 17.43% in sentence encoding tasks. Additionally, we investigate the impact of model initialization on convergence and evaluate the significance of domain-specific models in generating semantically meaningful sentence embeddings. The models are publicly available at: https://huggingface.co/crisistransformers.

Protracted development of stick tool use skills extends into adulthood in wild western chimpanzees.

Tool use is considered a driving force behind the evolution of brain expansion and prolonged juvenile dependency in the hominin lineage. However, it remains rare across animals, possibly due to inherent constraints related to manual dexterity and cognitive abilities. In our study, we investigated the ontogeny of tool use in chimpanzees (Pan troglodytes), a species known for its extensive and flexible tool use behavior. We observed 70 wild chimpanzees across all ages and analyzed 1,460 stick use events filmed in the Taï National Park, Côte d'Ivoire during the chimpanzee attempts to retrieve high-nutrient, but difficult-to-access, foods. We found that chimpanzees increasingly utilized hand grips employing more than 1 independent digit as they matured. Such hand grips emerged at the age of 2, became predominant and fully functional at the age of 6, and ubiquitous at the age of 15, enhancing task accuracy. Adults adjusted their hand grip based on the specific task at hand, favoring power grips for pounding actions and intermediate grips that combine power and precision, for others. Highly protracted development of suitable actions to acquire hidden (i.e., larvae) compared to non-hidden (i.e., nut kernel) food was evident, with adult skill levels achieved only after 15 years, suggesting a pronounced cognitive learning component to task success. The prolonged time required for cognitive assimilation compared to neuromotor control points to selection pressure favoring the retention of learning capacities into adulthood.

Особенности опознавания изображений фигур разного цвета и размера детьми 3-4 лет с использованием шумового фона

&lt;p&gt;The study is aimed at analyzing the training of early preschool children to recognize visual stimuli of different colors and sizes, based on a method we are developing, in an ordinary situation and with the introduction of acoustic interference, i.e., spoken noise presented through headphones. The materials of two empirical studies on a sample of 3&amp;ndash;4-year-old children (3.5&amp;plusmn;0.43) from a kindergarten in St. Petersburg, of whom there were 13 girls and 17 boys, are presented. The study was conducted with an interval of 2 months. The findings showed that children took longer to recognize and select stimuli of different colors and made more perseverative errors, compared to recognizing and selecting stimuli of different sizes. The introduction of acoustic noise significantly impaired task success, which was particularly pronounced when identifying irritants from colored stimuli. It is suggested that difficulties in identifying colored stimuli may be due to the absence of a sensory reference marked by a word, and the acoustic noise slowed sensorimotor response.&lt;/p&gt;

Bi-DexHands: Towards Human-Level Bimanual Dexterous Manipulation.

Achieving human-level dexterity in robotics remains a critical open problem. Even simple dexterous manipulation tasks pose significant difficulties due to the high number of degrees of freedom and the need for cooperation among heterogeneous agents (e.g., finger joints). While some researchers have utilized reinforcement learning (RL) to control a single hand in manipulating objects, tasks that require coordinated bimanual cooperation are still under-explored due to the fewer suitable environments, which can result in difficulties and sub-optimal performance. To address these challenges, we introduce Bi-DexHands, a simulator with two dexterous hands featuring 20 bimanual manipulation tasks and thousands of target objects, designed to match various levels of human motor skills based on cognitive science research. We developed Bi-DexHands in Issac Gym, enabling highly efficient RL training at over 30,000 frames per second using a single NVIDIA RTX 3090. Based on Bi-DexHands, we present a comprehensive evaluation of popular RL algorithms in different settings, including single-agent/multi-agent RL, offline RL, multi-task RL, and meta RL. Our findings show that on-policy algorithms, such as PPO, can master simple manipulation tasks that correspond to those of 48-month-old babies, such as catching a flying object or opening a bottle. Furthermore, multi-agent RL can improve the ability to perform manipulations that require skilled bimanual cooperation, such as lifting a pot or stacking blocks. Despite achieving success in individual tasks, current RL algorithms struggle to learn multiple manipulation skills in most multi-task and few-shot learning scenarios. This highlights the need for further research and development within the RL community.

UViT-Seg: An Efficient ViT and U-Net-Based Framework for Accurate Colorectal Polyp Segmentation in Colonoscopy and WCE Images.

Colorectal cancer (CRC) stands out as one of the most prevalent global cancers. The accurate localization of colorectal polyps in endoscopy images is pivotal for timely detection and removal, contributing significantly to CRC prevention. The manual analysis of images generated by gastrointestinal screening technologies poses a tedious task for doctors. Therefore, computer vision-assisted cancer detection could serve as an efficient tool for polyp segmentation. Numerous efforts have been dedicated to automating polyp localization, with the majority of studies relying on convolutional neural networks (CNNs) to learn features from polyp images. Despite their success in polyp segmentation tasks, CNNs exhibit significant limitations in precisely determining polyp location and shape due to their sole reliance on learning local features from images. While gastrointestinal images manifest significant variation in their features, encompassing both high- and low-level ones, a framework that combines the ability to learn both features of polyps is desired. This paper introduces UViT-Seg, a framework designed for polyp segmentation in gastrointestinal images. Operating on an encoder-decoder architecture, UViT-Seg employs two distinct feature extraction methods. A vision transformer in the encoder section captures long-range semantic information, while a CNN module, integrating squeeze-excitation and dual attention mechanisms, captures low-level features, focusing on critical image regions. Experimental evaluations conducted on five public datasets, including CVC clinic, ColonDB, Kvasir-SEG, ETIS LaribDB, and Kvasir Capsule-SEG, demonstrate UViT-Seg's effectiveness in polyp localization. To confirm its generalization performance, the model is tested on datasets not used in training. Benchmarking against common segmentation methods and state-of-the-art polyp segmentation approaches, the proposed model yields promising results. For instance, it achieves a mean Dice coefficient of 0.915 and a mean intersection over union of 0.902 on the CVC Colon dataset. Furthermore, UViT-Seg has the advantage of being efficient, requiring fewer computational resources for both training and testing. This feature positions it as an optimal choice for real-world deployment scenarios.

Contribution to pulmonary diseases diagnostic from X-ray images using innovative deep learning models

Pulmonary disease identification and characterization are among the most intriguing research topics of recent years since they require an accurate and prompt diagnosis. Although pulmonary radiography has helped in lung disease diagnosis, the interpretation of the radiographic image has always been a major concern for doctors and radiologists to reduce diagnosis errors. Due to their success in image classification and segmentation tasks, cutting-edge artificial intelligence techniques like machine learning (ML) and deep learning (DL) are widely encouraged to be applied in the field of diagnosing lung disorders and identifying them using medical images, particularly radiographic ones. For this end, the researchers are concurring to build systems based on these techniques in particular deep learning ones. In this paper, we proposed three deep-learning models that were trained to identify the presence of certain lung diseases using thoracic radiography. The first model, named “CovCXR-Net”, identifies the COVID-19 disease (two cases: COVID-19 or normal). The second model, named “MDCXR3-Net”, identifies the COVID-19 and pneumonia diseases (three cases: COVID-19, pneumonia, or normal), and the last model, named “MDCXR4-Net”, is destined to identify the COVID-19, pneumonia and the pulmonary opacity diseases (4 cases: COVID-19, pneumonia, pulmonary opacity or normal). These models have proven their superiority in comparison with the state-of-the-art models and reached an accuracy of 99,09 %, 97.74 %, and 90,37 % respectively with three benchmarks.