The integration of technology into cooking and eating routines often leads to distractions, hindering mindfulness. Mindful eating involves awareness of the present moment and attention to sensory experiences. While mindful eating has been promoted, mindful cooking remains underexplored. As conversational agents become increasingly popular in diverse fields like education, healthcare, and e-commerce, we are leveraging this technology to promote mindfulness practices that enhance the cooking and eating experience. This paper introduces MyndFood, a conversational agent designed to promote mindful cooking and eating. We describe the development process of MyndFood and present the outcomes from a user experiment involving 40 participants who engaged with the system while preparing sushi. The study employed a between-subjects design featuring two conditions for the agent: mindful and non-mindful. Results indicate that participants in the mindful condition experienced greater enjoyment and heightened food awareness compared to those in the non-mindful condition. These findings suggest that conversational agents can enhance hedonic experiences and improve food awareness by promoting mindful practices.

Deep neural networks (DNNs) are extensively used in both current and future manufacturing, transportation, and healthcare sectors. The widespread use of neural networks in highly safety-critical applications has made it necessary to prevent catastrophic issues from arising during prediction processes. In fact, misreading a traffic sign by an autonomous car or performing an incorrect analysis of medical records could put human lives in danger. With this awareness, the number of studies related to deep neural network verification has increased dramatically in recent years. In particular, formal guarantees regarding the behavior of a DNN under particular settings are provided by model checking, which is crucial in safety-critical applications where network output errors could have disastrous effects. Model checking is an effective approach for confirming that neural networks perform as planned by comparing them to clearly stated qualities. This paper aims to highlight the critical need for and present challenges associated with using model-checking verification techniques to verify deep neural networks before relying on them in real-world applications. It examines state-of-the-art research and draws the most prominent future directions in the model checking of neural networks.

IntroductionFacial expressions play a crucial role in human emotion recognition and social interaction. Prior research has highlighted the significance of the eyes and mouth in identifying emotions; however, limited studies have validated these claims using robust biometric evidence. This study investigates the prioritization of facial features during emotion recognition and introduces an optimized approach to landmark-based analysis, enhancing efficiency without compromising accuracy.MethodsA total of 30 participants were recruited to evaluate images depicting six emotions: anger, disgust, fear, neutrality, sadness, and happiness. Eye-tracking technology was utilized to record gaze patterns, identifying the specific facial regions participants focused on during emotion recognition. The collected data informed the development of a streamlined facial landmark model, reducing the complexity of traditional approaches while preserving essential information.ResultsThe findings confirmed a consistent prioritization of the eyes and mouth, with minimal attention allocated to other facial areas. Leveraging these insights, we designed a reduced landmark model that minimizes the conventional 68-point structure to just 24 critical points, maintaining recognition accuracy while significantly improving processing speed.DiscussionThe proposed model was evaluated using multiple classifiers, including Multi-Layer Perceptron (MLP), Random Decision Forest (RDF), and Support Vector Machine (SVM), demonstrating its robustness across various machine learning approaches. The optimized landmark selection reduces computational costs and enhances real-time emotion recognition applications. These results suggest that focusing on key facial features can improve the efficiency of biometric-based emotion recognition systems without sacrificing accuracy.

Early and accurate detection of plant diseases is crucial for making informed decisions to increase the yield and quality of crops through the decision of appropriate treatments. This study introduces an automated system for early disease detection in plants that enhanced a lightweight model based on the robust machine learning algorithm. In particular, we introduced a transformer module, a fusion of the SPP and C3TR modules, to synthesize features in various sizes and handle uneven input image sizes. The proposed model combined with transformer-based long-term dependency modeling and convolution-based visual feature extraction to improve object detection performance. To optimize a model to a lightweight version, we integrated the proposed transformer model with the Ghost module. Such an integration acted as regular convolutional layers that subsequently substituted for the original layers to cut computational costs. Furthermore, we adopted the SIoU loss function, a modified version of CIoU, applied to the YOLOv8s model, demonstrating a substantial improvement in accuracy. We implemented quantization to the YOLOv8 model using ONNX Runtime to enhance to facilitate real-time disease detection on strawberries. Through an experiment with our dataset, the proposed model demonstrated mAP@.5 characteristics of 80.30%, marking an 8% improvement compared to the original YOLOv8 model. In addition, the parameters and complexity were reduced to approximately one-third of the initial model. These findings demonstrate notable improvements in accuracy and complexity reduction, making it suitable for detecting strawberry diseases in diverse conditions.

IntroductionPansharpening is an important remote sensing task that aims to produce high-resolution multispectral (MS) images by combining low-resolution MS images with high-resolution panchromatic (PAN) images. Although deep learning-based pansharpening has shown impressive results, the majority of these models frequently struggle to balance spatial and spectral information, resulting in artifacts and a loss of detail in pansharpened images. Furthermore, these models may fail to properly integrate spatial and spectral information, leading to poor performance in complex scenarios. Additionally, these models face challenges such as gradient vanishing and overfitting.MethodsThis paper proposes a dual-path and multi-scale pansharpening network (DMPNet). It consists of three modules: the feature extraction module (FEM), the multi-scale adaptive attention fusion module (MSAAF), and the image reconstruction module (IRM). The FEM is designed with two paths, namely the primary and secondary paths. The primary path captures global spatial and spectral information using dilated convolutions, while the secondary path focuses on fine-grained details using shallow convolutions and attention-guided feature extraction. The MSAAF module adaptively combines spatial and spectral data across different scales, employing a self-calibrated attention (SCA) mechanism for dynamic weighting of local and global contexts and a spectral alignment network (SAN) to ensure spectral consistency. Finally, to achieve optimal spatial and spectral reconstruction, the IRM decomposes the fused features into low- and high-frequency components using discrete wavelet transform (DWT).ResultsThe proposed DMPNet outperforms competitive models in terms of ERGAS, SCC (WR), SCC (NR), PSNR, Q, QNR, and JQM by approximately 1.24%, 1.18%, 1.37%, 1.42%, 1.26%, 1.31%, and 1.23%, respectively.DiscussionExtensive experimental results and evaluations reveal that the DMPNet is more efficient and robust than competing pansharpening models.

IntroductionThe recognized credibility crisis in scientific research has led to an increasing focus on reproducibility studies, particularly in computer science. Existing studies predominantly examine specific technological aspects of reproducibility but neglect the critical interplay between authors and publishers in enabling reproducible computational scientific research.MethodsA systematic review was conducted following the PRISMA Literature Review methodology, complemented by a Journals Survey. This approach enabled a comprehensive analysis of reproducibility policies and software/data management practices in scientific computing journals.ResultsThe survey revealed significant variability in reproducibility policies and practices across computer science journals. Many gaps and challenges were identified, including inconsistencies in policy enforcement, lack of standardized tools, and insufficient recognition of software as a research artifact. The analysis highlighted the potential of Reproducibility as a Service (RaaS) as an innovative solution to address these challenges.DiscussionThis study underscores the need for improved standardization and implementation of reproducibility policies. Strategies to enhance reproducibility include fostering collaboration among authors, publishers, and technology providers, as well as recognizing software as a critical research output. The findings aim to guide stakeholders in bridging the current gaps and advancing the reproducibility of computational scientific articles.

IntroductionDistance learning has seen a significant increase as educational institutions have shifted toward offering online courses. Although some institutions quickly adapted, many struggled to modify traditional materials for online learners. Time was crucial for institutions lacking experience in remote teaching. Designing engaging online cybersecurity modules for diverse students is a major challenge. With the growing popularity of online courses, it is necessary to examine the teaching methods used. This paper presents a systematic literature review on the current state of online cybersecurity education at universities. Using the PRISMA approach, the study identifies prevalent themes and addresses key research questions. This study aims to analyze academic articles to highlight key findings on how universities teach cybersecurity courses online.MethodsThe authors conducted a systematic review of scholarly articles, adhering to the PRISMA approach for the period from January 2010 to August 2024. PRISMA offers a structured approach to planning, executing, and reporting systematic reviews in various fields, including healthcare and social sciences.ResultsThe review revealed several key findings on the design of online cybersecurity courses. Learner-centered approaches were commonly used, featuring active learning and practical applications. Effective instructional methods included collaborative learning, case studies, and simulations, which promoted student engagement and critical thinking. Universities emphasize practical skills evaluation and knowledge acquisition through project-based assessments. The role of IT tools was highlighted, with virtual laboratories, gamification, and simulations providing hands-on experiences, enhancing motivation, and facilitating active learning.DiscussionThis systematic review provides a comprehensive overview of the current online cybersecurity education practices in online universities. As a pioneering effort, it offers educators and curriculum developers valuable insight into designing effective online cybersecurity programmes to enhance teaching and learning practices. The review of online cybersecurity education highlighted several key findings. Learner-centered approaches, which incorporate active learning practices and practical applications, were prevalent. Effective instructional methods included collaborative learning, case studies, and simulations, which fostered student engagement and critical thinking. Assessments focused on the acquisition of practical skills and knowledge, using project-based tasks, practical exercises, and online quizzes. IT tools played a significant role, with virtual laboratories, gamification, and simulation environments that enhanced hands-on experiences, motivation, and active learning.

This perspective article details the program of sustainable computing workshops launched within the High Energy Physics department at Deutsches Elektronen-Synchrotron (DESY) in 2023. The workshop series targets scientific users of the National Analysis Facility, hosted at DESY, in order to promote sustainable software development and usage across a wide range of applications. Details of the structure of the workshops, as well as the reception amongst participants, will be presented. In addition, plans for the expansion of the workshop series will be outlined.

In this paper, we combine the deep image prior (DIP) framework with a style transfer (ST) technique to propose a novel approach (called DIP-ST) for image inpainting of artworks. We specifically tackle cases where the regions to fill in are large. Hence, part of the original painting is irremediably lost, and new content must be generated. In DIP-ST, a convolutional neural network processes the damaged image while a pretrained VGG network forces a style constraint to ensure that the inpainted regions maintain stylistic coherence with the original artwork. We evaluate our method performance to inpaint different artworks, and we compare DIP-ST to some state-of-the-art techniques. Our method provides more reliable solutions characterized by a higher fidelity to the original images, as confirmed by better values of quality assessment metrics. We also investigate the effectiveness of the style loss function in distinguishing between different artistic styles, and the results show that the style loss metric accurately measures artistic similarities and differences. Finally, despite the use of neural networks, DIP-ST does not require a dataset for training, making it particularly suited for art restoration where relevant datasets may be scarce.

Assistive technologies play a major role in bridging the accessibility gap in arts (especially paintings). Despite the constant advancements in these areas, the visually impaired often encounter challenges in independently experiencing and interpreting paintings. Our goal was to effectively convey the contents of a painting to visually impaired students using selected multi-sensory stimuli (tactile, auditory, and somatosensory) to compensate for the loss of input from the sense of sight. A prototype (named SEMA—Specially Enhanced Multi-sensory Art) was developed around a simple painting to incorporate descriptive outputs of the aforementioned stimuli. The prototype was developed and refined iteratively with the visually impaired students at the University of Colombo, Sri Lanka. We evaluated all the systems individually using the user feedback and several quantitative and qualitative measures were adopted to analyze the results. The final user study with 22 visually impaired participants yielded a strong preference (92.6 %) for the prototype and highlighted its potential to enhance the art experiences. The findings of this study contribute to the further exploration of multi-sensory integration in entertainment and its impact on the visually impaired community.