Health Behavior Change (HBC) applications hold significant promise for supporting healthier lifestyles, such as improving physical activity, diet, or sleep. However, designing effective digital interventions that lead to sustained behavior change remains a challenge for designers and industry professionals. This paper introduces VITAL, a framework that offers a more structured and efficient approach to creating impactful HBC applications. VITAL is built on three core components: (1) process motivators, (2) intervention techniques, and (3) the design layer. It helps translate behavioral strategy into design by linking emotional motivators to evidence-based techniques—including Behavior Change Techniques (BCTs), personalization, and gamification—and converting them into practical product features through structured design prompts and actionable solutions. In doing so, VITAL bridges the gap between theory and user experience.

This study explores the intuitiveness and user acceptance of gesture-based interaction paradigms within extended reality (XR) environments, using the Apple Vision Pro headset as a testing platform. Through a three-stage evaluation process—comprising Pre-Test, Mid-Test, and Post-Test questionnaires—data was collected from participants with varied levels of technological experience. The findings indicate a strong preference for naturalistic, embodied interactions such as hand gestures and voice input over traditional digital methods like keyboard typing. Notably, all participants successfully engaged with 3D models and immersive interface elements without prior instruction, highlighting the effectiveness of real-world gesture metaphors in facilitating user adaptation. Despite limited prior XR exposure, the majority of users found the system intuitive and enjoyable, especially when interacting with spatial content in three-dimensional formats. These results underscore the viability of gesture-driven interaction models in future XR systems and emphasize the importance of aligning interface design with users’ innate cognitive and motor behaviors. The paper concludes by advocating for interdisciplinary approaches to advance inclusive and intuitive XR interaction paradigms, thereby supporting broader societal integration of immersive technologies.

Integrating Mixed Reality (MR) technologies in higher education represents an opportunity to create smart-university campuses. This article explores the development of the Mixed Reality University Campus. This complex system combines Virtual Reality (VR), Digital Twin concepts, Natural Language Processing (NLP), ChatBots, and Interactive Materials to empower the academic experience at the Faculty of Computer Science in Iasi. Unlike conventional simulation-based applications, this project aims to fully replicate key student activities, ensuring an immersive, interactive, and highly accessible learning environment. By leveraging mixed reality solutions, the initiative provides a study platform that supports students, offering an alternative educational approach that proved essential during global disruptions such as the COVID-19 pandemic. The Mixed Reality University Campus represents an advancement in e-learning and human-computer interaction, creating the way for future innovations that will redefine how students interact in educational situations. This research demonstrates how such emerging technologies can create more engaging, efficient, and accessible academic experiences, shaping the future of higher education.

Recent advances in Large Language Models (LLMs) have created new opportunities for automated and efficient diagram generation. This study introduces DiagramAI, a tool designed for AI-driven diagram generation and editing. To evaluate its feasibility, we assess the capabilities of seven LLMs in generating five types of software diagrams: use case, class, architectural, sequence, and C4 context diagrams. We compare two-generation approaches, direct Mermaid code output, and structured YAML-based representations, analyzing their syntactic correctness, structural consistency, and adherence to diagramming conventions. The results reveal significant variations in model performance, with some LLMs demonstrating strong syntax adherence, while others struggle with diagram structure and semantic accuracy.

The process of creating a schedule requires a lot of patience, time, and work. A timetable is made for a number of reasons, such as scheduling lectures at universities and schools, making timetables for bus and train schedules, and many more. A timetable takes a lot of time and labor to develop. This article provides a thorough analysis of the planning and execution of an application for creating timetables that can manage a variety of characteristics, including classes, students, teachers, subjects, resources, and classrooms. The program simplifies the preparation of timetables by utilizing a well-organized database and offers an intuitive user interface for effective maintenance. Aspect-Oriented Programming (AOP) and Monitoring-Oriented Programming (MOP) approaches are integrated into the system to optimize a variety of scheduling factors and provide a balanced and ideal schedule. Moreover, the application offers synchronization capabilities with personal calendars, facilitating enhanced coordination for teachers and students. This synchronization ensures that the generated timetable aligns with individual commitments, thus fostering improved time management.

This study assesses the diagnostic capacities of four top ranking AI chatbots in Romanian medical letter interpretation: Gemini, Copilot, DeepSeek, and Qwen. Extending earlier studies on AI interpretation of Holter investigation images, these tools’ capacity to produce credible diagnoses from medical texts, with and without images from Holter investigations, were analyzed. In order to rate diagnostic predictions, a weighted scoring system was put in place and performed five repetitions per instrument version and case for a more rigorous approach. The findings show that all four AI systems can identify key diagnostic elements that often match the real medical diagnoses, though with varying degrees of accuracy and consistency ranging from 81.2% to 92.44%, with Qwen achieving the highest consistency after incorporating imaging data. Response times varied significantly (31-84 seconds), with Gemini demonstrating the fastest average response. The study highlighted the importance of prompt engineering, as structured prompts with clear instructions produced more organized and relevant responses. Some limitations were identified, such as occasionally providing contradictory diagnoses and not always recognizing certain diseases. These findings suggest there is no universally superior AI tool for cardiology diagnosis, as optimal performance depends on the specific clinical presentation and input format. The multidisciplinary team, which included a cardiologist and computer scientists, made review possible from two points of view: medical accuracy and consistency across multiple test cases. The research aims to provide valuable insights for healthcare professionals regarding the promising capabilities and current limitations of freely available AI systems in medical contexts.

XR development is still time-consuming due to the complexity of creating proper 3D assets and real-time performance requirements. From the whole pipeline of development, the most time-consuming part is still 3D data acquisition. An alternative to hand-made assets is 3D scanning, but there are many options, from cheap, consumer-grade devices to high-end professional-grade equipment that can cost even hundreds of thousands of dollars. Each project has different needs. Our article explores a series of low-to-medium budget pipelines in the context of XR development, wishing to help the reader identify the best solution for their specific project. We compare both technically and qualitatively nine popular techniques, inside indoor environments, and came up with a list of conclusions based on both what to use now, and what to expect in the future.

In the emerging context of immersive technologies, using Virtual Reality (VR) in education provides innovative ways to experience and understand complex concepts such as chemical reactions. This project demonstrates how VR can revolutionize how students interact with reactants and reaction products, allowing them to observe and manipulate chemical reactions in a controlled and fully immersive environment. By creating a virtual laboratory, participants can directly experience the effects and dynamics of various chemical reactions without the risks of handling real substances. This approach not only increases student engagement and curiosity but also improves information retention and a deep understanding of chemical processes. This project illustrates how the combination of Artificial Intelligence and Virtual Reality can transform STEAM education by providing a more interactive, safe, and accessible educational method. Our solutions present significant advances in educational practices, with the technologies used increasing student engagement exponentially. In other words, by incorporating Artificial Intelligence algorithms within virtual environments, we can take the field of eLearning to a higher level.

The profession of Medical Bioengineer is a complex one because it is an interdisciplinary job, on the border between medicine and engineering. The professional training of students within the Medical Bioengineering specialization, especially the clinical engineering branch, requires knowledge both from a technical and functional point of view of all medical equipment in health facilities. In this study, we propose the development of a virtual reality platform adapted to biomedical training in the field of medical devices. It immerses students in realistic medical scenarios based on medical devices, integrating feedback and physical simulations for a dynamic learning environment. The platform’s advanced grading system accurately evaluates student interactions, considering task accuracy, technical finesse, and efficiency. It includes a robust progress-tracking component that allows educators to monitor individual progress over time, providing valuable insights into student learning trajectories. This project bridges the gap between traditional medical training and the revolutionary potential of VR. It aims to provide an immersive learning experience where students can interact with medical instruments and devices in a simulated environment. The application’s system and progress tracking feature provide an objective assessment of comprehension abilities, enhancing the adaptability of medical education. The resulting VR platform greatly enhances the educational experience and aligns with the faculty’s vision of progressive educational methodologies, as well as the standards for learning in VR [7]. Because in some situations, during the years of study, students cannot have access to all types of medical equipment, this platform allows them direct interaction with a wide range of medical equipment.

In the dynamic landscape of digital interactions, the concept of Digital Twins is shaping the future of various applications. This article explores the use case of Digital Twins, which are virtual models of physical objects or systems, in conjunction with emerging technologies such as Extended Realities (XR). By using Digital Twins, we can create comprehensive simulations that enhance real-world processes, leading to greater efficiency and innovation across multiple areas. This paper examines the impact of Digital Twins on digitalized institutions, illustrating how these technologies improve decision-making, predictive maintenance, and immersive user experiences. We highlight several use cases where the integration of Digital Twins with VR, AR, and AI has significantly enhanced operational capabilities and user engagement. The synergy of these advanced technologies marks a new era of digital transformation, setting the stage for smarter, more responsive, and interconnected systems in the digital future.

In today’s complex IT job market, young students and graduates often find themselves overwhelmed by the multitude of options and challenges they face. However, this platform offers a unique and indispensable solution to this problem. Recognizing the need to simplify the process of finding and accessing employment opportunities and internships, we have developed and implemented an application that addresses these concerns. By consolidating and centralizing job offers and internships from the most prestigious recruitment platforms in Romania, we have created a user-friendly and intuitive environment where career opportunities are readily available. This platform not only streamlines the employment search but also facilitates the transition from academia to the professional world, equipping users with the necessary tools to confidently navigate their path to success in the IT field. During the development of the application, we utilized a web scraping technique to gather information on all available jobs from the major recruitment platforms in Romania. This initial step was followed by a meticulous data processing process, ensuring that the information was stored accurately and efficiently in the application’s database. One of the challenges we encountered was handling job listings that did not include salary information. To address this issue and provide users with comprehensive information, we implemented a salary estimation method for these jobs. This estimation was based on data from similar positions, ensuring that users have access to the most up-to-date and relevant information when making decisions about their IT careers.

In the evolving landscape of digital interaction, the integration of Artificial Intelligence (AI) and Mixed Realities (MR) presents unprecedented opportunities to enhance learning environments. This article explores the potential of AI-driven tools and techniques and MR platforms to transform traditional educational models, focusing on applications applied to student needs. By incorporating advanced AI algorithms and immersive MR experiences, we can offer personalized learning pathways and interactive, engaging content that meets the diverse needs of students across various disciplines. Our solutions highlight multiple use cases where these technologies have significantly increased student engagement and learning outcomes. The combination of Artificial Intelligence and Mixed Realities not only levels the playing field for access to high-quality education/training while also creating new paths in the eLearning field, showcasing significant progress in educational practices and accessibility.

This paper explores the practical implementation of 3D Convolutional Neural Networks (CNNs) for real-time violence detection in surveillance scenarios, focusing on applications such as traffic aggression and bullying detection. In recent years, the proliferation of surveillance cameras has provided a vast amount of visual data, necessitating efficient and accurate automated methods for threat identification. In this study, we investigate the feasibility and effectiveness of employing 3D CNNs for violence detection in real-life scenarios. We propose a comprehensive framework that integrates data augmentation techniques and fine-tuning strategies to address the challenges of limited annotated data and diverse environmental conditions. Furthermore, we conduct extensive experiments on benchmark datasets to evaluate the performance of the proposed approach in detecting physical altercations. In summary, this study highlights the efficacy of 3D CNNs in detecting physical altercations, thereby contributing to the advancement of automated surveillance systems aimed at enhancing public safety. The findings presented herein not only underscore the practical utility of deep learning techniques in addressing real-world challenges but also point towards future avenues for improving the accuracy and reliability of violence detection algorithms through multimodal integration.

This paper explores advanced music generation through hybrid models combining deep neural networks, machine learning algorithms, variational autoencoders (VAEs), long short-term memory (LSTM) networks, and Transformers to create diverse and engaging musical experiences. Our research aims to advance the understanding of music’s impact on our lives and develop methodologies to create diverse and engaging musical experiences tailored to individual preferences. We begin by extracting relevant features from a large and diverse collection of music samples from different genres. These features, encompassing spectral properties, rhythmic patterns, and tonal characteristics, serve as the foundation for our generation models. To generate music, we explore the potential of VAEs, LSTMs, and Transformers, each offering unique capabilities for handling different aspects of the task. VAEs are employed to learn a continuous latent space representation of the music samples, enabling the generation of novel compositions within a specified genre. LSTMs and Transformers, on the other hand, are used to model the temporal dependencies and intricate patterns inherent in music. While not claiming state-of-the-art performance, our approach demonstrates promising outcomes in generation tasks, showcasing its potential to enhance music-related applications such as recommendation systems and creative tools for composers.

In recent years, the way we make the decision to read a book has diversified a lot. From simple methods based on friends’ recommendations or the desire to read a book after watching a movie based on that book, to more complex methods that use artificial intelligence algorithms that build increasingly advanced user profiles. In this context, our project comes with a solution in virtual reality, where the user receives various types of information about books (textual, visual, auditory) and recommendations based on the advanced profiles we build for them. Thus, he can make the decision to read a book based on more information that he accesses within an interesting experience given by the virtual space where he interacts with them. The experiments carried out showed the readers’ interest and desire to explore other methods different from the traditional ones to find the desired book. They appreciated the fact that different types of information can be brought in the same application and can contribute to their final decision in choosing the desired book.

For a long time, people have been finding themselves in need of a way to ease the process of obtaining medical care and papers. Getting a diagnosis or treatment required going to a medic, and for various reasons (e.g., proximity, health condition) this visit would be difficult or impossible. Recently, several solutions for identifying a disease or getting medication recommendations through online methods have been created. In our application, users can log in their symptoms and get a diagnosis that will be checked by a doctor and approved when accurate, and also a recommended treatment, without having to go in person to the doctor’s office. For this, we used existing data sets and advanced machine-learning techniques. The results obtained are promising and make us hope that such solutions will be used on a large scale in the future.

Documents have been used across history ever since civilized societies first began appearing. Documents are used everywhere today in our daily activities and were affected by technological leaps. From documents written on paper, we switched to digital documents. One of the technological fields that are dealing with documents is Computer Vision, specifically OCR, or optical character recognition. OCR is the process in which an image containing text is converted into digital text format [1]. Because computers are used everywhere nowadays, systems have already been designed for working with documents. In many systems that deal with documents, there is still a need for manual work. This paper proposes a way in which OCR can be applied to official documents for the extraction of their text.

In order to create a system for Image Semantic Segmentation that works similarly to the human way of approaching visual problems, we decided to use divide-and-conquer to break an image into smaller pieces, called microimages. We use Spectral Clustering to understand the features and group them to identify the location of different objects in the image. We trained a Convolutional Neural Network to compute the affinity matrix required by Spectral Algorithm. We managed to obtain competitive results on the Berkeley Segmentation Dataset and we hope to continue to improve the method.

In recent years, artificial intelligence (AI) has started to be used more and more in the medical field. This paper presents a study focused on malaria classification based on segmented blood cells from images collected from the National Institutes of Health (NIH) database. The research involved the development of a handcrafted convolutional neural network (CNN), as well as experimentation with various fine-tuning approaches using the VGG16 architecture. The conducted experiments have yielded promising results, providing empirical evidence for the potential effectiveness of these techniques in future applications. The augmented CNN achieved an impressive accuracy of 96.51%, while the VGG16 fully trainable model outperformed it with an accuracy of 96.69%. A problem that needs to be analyzed more carefully in the future concerns the explainability of the results so that they can be used with confidence by healthcare professionals.

Alzheimer's disease (AD) is a neurodegenerative disorder that progressively affects cognitive function and which is the leading cause of dementia. Early diagnosis of AD is critical as it allows for timely intervention, and improves the individual's quality of life, having the potential to reduce the burden of the disease on individuals, families, and society. This study presents AlzDiagnostics, a mobile application whose primary goal is to offer a user-friendly tool for the early diagnosis of AD, integrating established clinical tools, state-of-the-art machine learning techniques, and multiple diagnostic approaches. Furthermore, the application seeks to assess the effectiveness and accuracy of the Mini-Mental State Examination (MMSE) test, machine learning models trained on MRI images, as well as ones trained on clinical data, evaluating their contributions to enhance early diagnosis. Choosing the best classification models for the diagnosis module involved the conduct of numerous classification experiments on an extensive MRI dataset obtained from Kaggle, which consisted of 6400 MRI images from different sources including websites, hospitals, and public repositories. These experiments encompassed the exploration of various parameter configurations, leading to the identification of a top-performing model that exhibited an accuracy rate of 98%. Additionally, extensive experiments were conducted on clinical data retrieved from the Alzheimer's Disease Neuroimaging Initiative (ADNI) dataset, which comprised entries from 2425 patients. By incorporating features such as gender, marital status, education, race, and MMSE scores, a second machine learning model was developed, yielding an accuracy level of 76%.

The technological progress in computer science (particularly, in machine learning) has contributed to the improvement of medical services, both in detecting and treating diseases. The large volumes of data, that are overwhelming for human experts (doctors, nurses), can easily be managed by automated systems, as long as we have the computational resources. Obviously, human experts are still essential in the process - we think of the use of computer science in medicine as a collaboration between medical staff and artificial intelligence. The usual types of data that can be processed by automated systems are text, sound, and image types. In this paper, we approach the diagnosis subject and focus on data consisting of sound. We created a heart murmur detection system - it analyzes recordings and tells the user whether the sound samples indicate a heart murmur or not, based on a trained machine learning model. One of the main advantages of our system is the fact that we ran a large number of experiments, with different configurations of denoising techniques and features taken into consideration. We were able to draw some interesting conclusions, for example, we found out which features are the most important for the classification and which features are not worth computing. Also, our work denotes a thorough understanding of sound processing.

Since the release of blockchain technology back in 2008, a huge number of blockchain-based systems have been developed. As of Sep’22, there are at least 1,000 blockchain systems ready to be used, with at least four types of networks: public, private, hybrid, and consortium. With the many options you can choose from, a valid question arises: How can we choose a system that fits best with our needs? The purpose of Blockchain Pedia is to help users make the best decision in choosing the right blockchain system, giving the possibility to see the differences between multiple options. The application is designed for anyone who wants to build a project based on blockchain technology, giving recommendations based on the needs of the developer.

The early detection of Alzheimer's disease is a significant research priority in the field, as it enables timely medical intervention and treatment that may potentially slow down the progression of the disease and improve the patient's quality of life. Research has shown that early indicators of this condition can be identified up to 20 years before symptoms onset through the use of Magnetic Resonance Imaging (MRI) scans. This study evaluates the potential of transfer learning and deep learning algorithms for the accurate diagnosis of Alzheimer's Disease (AD) using MRI scans. The evaluation focuses on two pre-trained models, ResNet-152 and AlexNet, using original vendor data from the ADNI dataset, as well as data subjected to skull stripping. The results showed that the ResNet-152 model achieved the highest accuracy of 99.96% on sagittal plane slices extracted from the original dataset. Furthermore, the study highlights that the models trained on original MRI images outperformed those trained on images that underwent skull stripping.

This paper presents a novel approach for detecting and finding similar clothing articles in images by combining structural shape detection and texture comparison. We utilized the Mask R-CNN model for structural shape detection and explored color hashing and Local Binary Pattern (LBP) histogram for texture comparison. Experiments were conducted using the DeepFashion2 dataset, demonstrating varying performance across different clothing articles and texture comparison algorithms. While the proposed solution achieved good detection rates for certain clothing articles, it faced challenges in detecting others due to the unequal distribution of articles in the dataset and limitations in texture comparison.

Computer science has an impact on every field of activity, including medicine. On one hand, in medicine, automated systems can detect several diseases. On the other hand, they are also involved in the treatment process. Specialized systems can recommend medication, taking into consideration the side effects, give lifestyle suggestions (e.g., sleep, physical exercises), and even indicate to the patients which medics should they go to. One of the main advantages of the automated systems is the ability to process large volumes of data, at higher rates than the human experts. The data may be represented as text, images, or sounds. In this paper, we approach the treatment subject, and focus on data represented as text. We created a system based on sentiment analysis (powered by natural language processing), that provides valuable information regarding conditions/drugs to the user, obtained from the reviews given by other patients.

In the recent period, neural networks are used in more and more applications and services across a broader and broader spectrum of industries and domains. One of the applications of neural networks is in artistic content generation, ranging from schematics and improving 3D rendering to AI-powered up-scaling of video-games through DLSS, and image and music generation. In this project, we tried various music generation methods in order to see their limitations, unrefined results, and the most popular generation of audio files.

Millions of people use Twitter every month, which makes it one of the most popular social networks worldwide. Currently, there is an enormous scope market with the potential to be optimized to increase Twitter posts’ popularity and engagement. In this paper, we present a method of predicting the number of likes a given post will receive. We introduce a deep learning model and training procedure that uses both computer vision and natural language processing to reach high accuracy when shown new data. Considered use-cases will show the situations in which our system behaves well and the situations in which we do not yet have a solution to improve the current results.

As a consequence of the day-to-day increase in energy requirements worldwide, many countries around the world are focusing on implementing renewable energy sources (RES) to become energy independent. Although this concern is usually addressed at a national level, we believe that it also has to be addressed at the individual level, i.e. the consumer segment has to contribute to the shift to a sustainable energy future. This paper aims to provide a tool that helps consumers evaluate potential investment opportunities in renewable energy solutions. The tool focuses on two forms of RES, namely solar energy and wind energy. Several comparison tests were performed on regions from Romania. However, the tool can be also extended to other regions around the world.

Atrial fibrillation is the most common sustained heart rhythm abnormality in clinical practice that can lead to well-known medical complications associated with increased mortality. The diagnosis of atrial fibrillation can be detected by using a short electrocardiogram recording (ECG). However, because the heartbeat is irregular and does not constantly present on a simple electrocardiogram, a single diagram is not enough for a final and certain diagnosis. To monitorizing one single patient requires hours of monitoring, important costs, and low yield. Our aim is to develop a rapid, inexpensive way to identify patients with atrial fibrillation using neural networks. This study serves to help the clinician with an automatic approach to give a quick and safe diagnosis for each patient population. The experiments show that this approach offers a promising atrial fibrillation classification and outperforms recently published studies that either use extracted features or raw data separately.

Nowadays anyone possesses at least one personal document, whether it is an identity card or a driving license. In many of our daily activities, working with documents or the necessity to present them in different contexts became something normal. Given the current situation, we are all going through, having to deal with the current pandemic situation, we have adapted to some extent to new ways of communicating, working, and resolving things, in general, using Internet tools and platforms. In this context, a small part of the fight we have with the bureaucracy has been won, many of the documents that once could only be brought only in physical locations can now be scanned and sent by email or attached to a form on the website. Hoping that the pandemic has also taught us good things, we want to believe that these ways of communication will be kept in place even after this issue is over, to avoid unnecessary waste of time and congestion. The focus of this project is developing an application that can create textual information data from a simple image provided by the user that will enable the possibility of sharing digital versions of documents, making our lives so much better.

Data privacy in photos is an important subject that is a major concern in today's world, especially given the compromising data it can contain or the rate at which the information can flow through the Internet. In this paper, we are proposing an approach that can mitigate the danger of sharing things like documents or personal images on the Internet. Our solution to this problem would be to obfuscate these images and documents with a set of randomly generated steps that would alter their content until the owner wants to access them again. Initially, the proposed solution can automatically identify faces, and has support for custom obfuscation on multiple defined areas. After that, it offers the possibility to export the keys separately for each of these areas, allowing at the time of deobfuscation access only to the desired areas.

Starting with everyday activities such as taking a walk at the outside environments, use of different public transportation methods or even buying different goods, visually impaired people need to put in a considerable effort to complete these so-called trivial tasks for sighted people. Given this, according to WHO (World Health Organization) it is estimated that there are at least 2.2 billion people who have a near or distance vision impairment. This paper presents an ongoing project that aims to address mainly two problems: the creation of a system designed to safely moving visually impaired people to points of interest using public transportation and designing of a communication mechanism that will balance between keeping the individual as safe as possible and also providing relevant indications. The current paper is divided in two main parts. In the first part, in addition to the study of similar systems, the factors underlying the architectural decisions on which the developed system was built, are debated and analysed. The second part presents the results obtained after a series of preliminary tests carried out with the help of visually impaired people. The application seems promising, being enthusiastically received by those who tested it, because it offers them safety and independence when travelling through the city.

Over time, buildings such as residential, historic or private start to deteriorate until they can endanger residents or those passing by. Such consequences, of the passage of time, also happen to streets, bridges or other built structures. This problem is difficult to identify even by authorized persons. At high altitudes, the specialist who assesses the condition of a building may be endangered or forced to endure adverse weather conditions. By automating computer systems, we can improve and monitor the process of identifying these cracks through applications that detect such features. The experiments of our project are based on U-Net, Autoencoder and VGG16 which show promising results.

Growth in social media platforms over the years have facilitated an enhancement in human communication. Platforms such as Facebook and Twitter are most ever-present in our lives and influencing how we speak, think, act and interact. The growth of fake news greatly impacts this phenomenon as it lowers one’s trust in the content presented. Dangerous is also the fact that readers can be behaviorally and psychologically profiled in order to be served specially crafted content with the intention of changing one’s opinion an action. One such is example is related to the 2016 U.S. presidential election campaign where fake news was a deciding factor in tipping the balance of power and outcome. It is hence of critical importance to develop tools that detect and combat such destructive content. This paper does not focus on the fake news detection problem but on a related problem regarding the prefiltration of the content sent to detection. Considering social media platforms have such diverse content and news is but a fraction of the data handled in the network it would be futile to label every post; not only it would not provide any value to the user, but it would also load the servers in a senseless manner. This work presents and evaluates a metric which is a score, from 0 to 100 of how close a text is to being a news content. The algorithm calculates a relevance score to assess if the content is news as well as uses information about the post and user (how many followers has, how many likes the tweet has and so on).

Temperature variations, climatic changes or human interventions are the main cause buildings and streets deteriorate over time. Since manual verification of such structures completely depends on a specialist’s knowledge and experience it is often time-consuming, costly, and dangerous, therefore an automated approach would be very beneficial. Automated devices using computer vision that detect anomalies in targeted structures have the potential of creating a better and safer environment, periodically checking for possible faults and improving worker safety. The system presented in this paper is an Android application with the purpose of field use, the goal being that either the worker could use it to detect potential hazards or an autonomous system might use it in inhospitable scenarios. The model created contains two modules, the U-Net algorithm, and the Android application; the novelty of the approach comes from offline use and accuracy, optimizations have been made in order for a reasonably powerful device to run the model and classify in real-time without the use of servers or mobile data. Adapting the neural network model to the mobile application was a challenge for our team, but in the end, the mobile application works as we intended with good results.

Healthcare is a vital human need and always had an important role in our lives, especially in the last year when the pandemic context raised a call for innovative and safe delivery methods. Together with the enlargement in the use of wearable sensors and smartphones, remote healthcare monitoring applications have unfolded at a fast pace, helping with the prevention of spreading diseases as well as to strengthen, help, and assist the health of a patient when a doctor is not physically available. This paper proposes a remote patient monitoring system within an application whose purpose is to bring patients closer to vital medical care. Based on voice interaction and wearable sensors, our application congregates the health data through the Google Fit app from the patient’s phone, and then it conducts analysis and sends automatic data alerts for detected anomalies, providing an appealing interface and capabilities to make it engaging to all age groups. Moreover, every patient is under the supervision of a doctor who is constantly monitoring the patient’s profile, ensuring assistance through presumptive diagnosis, treatments, or pieces of advice for boosting the quality of lifestyle.

In the time on structured surfaces (walls, roofs, bridges, streets, etc.) cracks appear and influence from the aesthetic point of view, but also from the point of view of their resistance and quality. Traditionally, crack detection is performed by human visual inspection, which is dangerous (when they need to climb buildings), subjective (depending on their experience in detecting the severity of a crack), and time-consuming (if we consider hundreds of buildings). Increasingly, artificial intelligence continues to evolve and we can use it to improve human performance and automate the process of crack detection. In order to improve this problem, we present an application that detects cracks in buildings that are difficult to access or would endanger human life. The architecture of our application is based on Convolutional Neural Network. In this paper, three different approaches are described and compared.

The fast rhythm of our lives makes it hard for us to find time to cook. When we do, we often have to choose between the commodity of cooking the same familiar food several times and the uncertainty of trying something new. This paper introduces AREasyCooking, an application which uses new technologies to help a user make healthier decisions about their culinary habits. Thus, using augmented reality and bar code reading, AREasyCooking helps users bring in the application the ingredients they have available. The application uses two major sources of recipes (well-known cooks and traditional recipes) to offer users possible variants of dishes with available ingredients. Finally, AREasyCooking guides users step by step through videos in preparing the recipe. Controlling these movies is done with the help of the eyes and the voice since hands are busy cooking.

Augmented Reality (AR) has been accepted as an effective learning method which means that it becomes complementary to traditional learning, especially in chemistry. In fact, AR is an interactive experience of a real-world environment. Before recent releases of cheap and affordable smart devices, AR large-scale applications in education were almost impossible. After a brief analysis of current trends in the use of AR, we propose a new system, named ARChemistry Learning, to support the Romanian educational system. In this study, we propose a modern AR tool in the chemistry education used to support children or anyone who wants to learn chemistry, to develop logic, and to explore the world seen only on a smart device. The purpose of this research is to demonstrate how effective these AR applications are, even that in Romania they are still in a pioneering phase.

Nowadays, social media credibility is a pressing issue for each of us who are living in an altered online landscape. The speed of news diffusion is striking. Given the popularity of social networks, more and more users began posting pictures, information, and news about personal life. At the same time, they started to use all this information to get informed about what their friends do or what is happening in the world, many of them arousing much suspicion. The problem we are currently experiencing is that we do not currently have an automatic method of figuring out in real-time which news or which users are credible and which are not, what is false or what is true on the Internet. The goal of this is to analyze Twitter in real-time using neural networks in order to provide us key elements about both the credibility of tweets and users who posted them. Thus, we make a real-time heatmap using information gathered from users to create overall images of the areas from which this fake news comes.

One of the industry’s most important trends in enterprise architecture is related to the use of microservices, to the detriment of monolithic architectures, which are beginning to no longer be used. Due the cloud-native architectures the deployment of microservices systems is more productive, flexible and cost effective. Anyway, a lot of companies started to migrate from one type of architecture to another, but it is still in the early phase. In this paper we address the challenges raised by the need to develop a scalable and fault tolerance system based on microservices. In our experiments we consider two types of microservices, simple and extended and the proposed solution proves to be an innovative one especially based on its dynamic behavior.

Visually challenged people may experience certain difficulties in their daily interaction with technology. That is essentially because the main way to exchange and process information is by written text, images or videos. Since the basic purpose of innovation is to improve people’s lifestyle, in this paper we propose a system that can make technology accessible to a broader group. Our prototype is presented as a mobile application based on vocal interaction, which can help people facing visual disorders consult their personal agenda, create an event, invite other friends to attend it, check the weather in certain areas and many other day-to-day tasks. Regarding the implementation, the project consists of a mobile application that interacts with a cloud based system, which makes it reliable and low in latency due to the resource availability in multiple global regions, provided by the newly emerging platform used in building the infrastructure. The novelty of the system lays in the highly flexible serverless architecture [1] that is open to extension and closed to modification through the set of autonomous cloud processing methods that sustain the base of the functionality. This distributed processing approach guarantees that the user always receives a response from his personal assistant, either by using artificial intelligence context generated phrases, by real-time cloud function processing or by fallback to the training answers.

This paper presents a general culture game, called Bob, implemented as a skill for Amazon’s software assistant Alexa. The main motivation of this work is to enable learning through games and smart devices, which are nowadays part of most children’s lives and homes. The Bob game provide users with general culture questions from the geography field, more specifically related to cities, countries, and lakes. The questions are formed by extracting and analyzing information from two major knowledge sources, DBpedia and Wikidata, both massively used in the natural language processing field. Bob can initiate general culture tests, during which questions will be automatically adapted to the knowledge level of the player through Computer Adaptive Testing (CAT). Following user specified settings, Bob also offers the possibility to rank players according to their performance and to post Twitter statistics on request. One major advantage of our solution comes from the fact that the application can be used through Amazon Echo or through a smartphone, allowing the child to access it from home or from any place with Internet connection.

In the last years big social networks like Facebook or Twitter admit that on their networks are fake and duplicate accounts, fake news and fake likes. With these accounts, their creators can distribute false information, support or attack an idea, a product, or an election candidate, influencing real network users in making a decision. In this paper, we present our system build with the aim of identifying fake users and fake news in the Twitter social network.

Over the last years, augmented reality was used in various domains, from medical, industrial design, modeling and production, robot teleoperation, military, entertainment, leisure activities to translation, facial recognition, assistance while driving, interior and exterior design, virtual friends, internet of things and eLearning. In eLearning, the combination between classical and augmented content (the later coming with 3D models, images, sounds, animations, Internet browsing, etc.) can help the teacher to better explain the content of the courses. In this paper, we present four augmented reality applications, created with the aim to improve communication and collaboration skills (two of them) and to ease the learning of biology and geography (the other two). The motivation behind these applications is to enhance the attractiveness of the classes, allow students to retrain new information more easily and reduce the stress behind tests when presented as games.

The need of humans to socialize and share information has led to a constantly growing Web, which has become a support for social media. Every day, worldwide users are pushing multimedia data towards their family, friends and the world at large. This is the reason why, web search has also become the main method for people to fulfill their information needs. The common modality used for image search on the web is based on text, the assumption being that the tags and the textual descriptions associated with photos are powerful ways to describe and retrieve images. The results are usually obtained by simply matching the terms of the query to an index of terms associated to the images in a corpus. The efficiency of this technique depends strongly on the tags associated to pictures, as well as their accuracy. Trying to search for images with bridges in this kind of systems, the result set will only contain images explicitly annotated with this term, but will fail to include images with Pont Neuf or Ponte Vecchio, if their tags do not contain the noun bridge. In this paper, we present a system designed to perform diversification in an image retrieval system, using semantic resources like YAGO, Wikipedia and WordNet.

Given the poor convergence of multi-objective evolutionary algorithms (MOEAs) demonstrated in several studies that address many-objective optimization, we propose a simple objective sampling scheme that can be incorporated in any MOEA in order to enhance its convergence towards the Pareto front. An unsupervised clustering algorithm is applied in the space of objectives at various moments during the search process performed by the MOEA, and only representative objectives are used to guide the optimizer towards the Pareto front during next iterations. The effectiveness of the approach is experimentally demonstrated in the context of the NSGA-II optimizer. The redundant objectives are eliminated during search when the number of clusters (representative objectives) is automatically selected by an unsupervised standard procedure, popular in the field of unsupervised machine learning. Furthermore, if after eliminating all the redundant objectives the number of conflicting objectives is still high, continuing to eliminate objectives by imposing a lower number of clusters speeds-up the convergence towards the Pareto front.