This research paper focuses on analog computers and creating a modular low-cost analog computer system in a single board computer form factor. The single-board analog computer will have the capacity to solve second-order differential equations. The capabilities and possibilities of the single board Analog computer will be explored as well as analog computing in general. The paper follows design science research methodology (DSRM) with the goal of creating and evaluating a working artifact. The artifacts' functionality is evaluated based on a demonstration of its ability to solve Mathieu’s differential equation as well as simulate a spring-mass dampening system. This paper proves that it is possible to create a low-cost analog computer in a modern form factor. The artifact is also placed in a larger contextual setting based on the empirical material provided where its value of it in a digital society is presented. For the world to continue its progression in computational power, but still, limit the already high energy usage, a drastic change is needed. This paper suggests adapting to analog/hybrid technology. To further the progression of analog/hybrid technology it must be made accessible to a wider group of people compared to today. The artifact in this paper offers a solution to this. 

The integration between cloud computing and vehicular ad hoc networks, namely, vehicular clouds (VCs), has become a significant research area. This integration was proposed to accelerate the adoption of intelligent transportation systems. The trustworthiness in VCs is expected to carry more computing capabilities that manage large-scale collected data. This trend requires a security evaluation framework that ensures data privacy protection, integrity of information, and availability of resources. To the best of our knowledge, this is the first study that proposes a robust trustworthiness evaluation of vehicular cloud for security criteria evaluation and selection. This article proposes three-level security features in order to develop effectiveness and trustworthiness in VCs. To assess and evaluate these security features, our evaluation framework consists of three main interconnected components: 1) an aggregation of the security evaluation values of the security criteria for each level; 2) a fuzzy multicriteria decision-making algorithm; and 3) a simple additive weight associated with the importance-performance analysis and performance rate to visualize the framework findings. The evaluation results of the security criteria based on the average performance rate and global weight suggest that data residency, data privacy, and data ownership are the most pressing challenges in assessing data protection in a VC environment. Overall, this article paves the way for a secure VC using an evaluation of effective security features and underscores directions and challenges facing the VC community. This article sheds light on the importance of security by design, emphasizing multiple layers of security when implementing industrial VCs.

The integration of the Internet of Things (IoT) and Machine Learning (ML) technologies has opened up for the development of novel types of systems and services. Federated Learning (FL) has enabled the systems to collaboratively train their ML models while preserving the privacy of the data collected by their IoT devices and objects. Several FL frameworks have been developed, however, they do not enable FL in open, distributed, and heterogeneous IoT environments. Specifically, they do not support systems that collect similar data to dynamically discover each other, communicate, and negotiate about the training terms (e.g., accuracy, communication latency, and cost). Towards bridging this gap, we propose ART4FL, an end-to-end framework that enables FL in open IoT settings. The framework enables systems' users to configure agents that participate in FL on their behalf. Those agents negotiate and make commitments (i.e., contractual agreements) to dynamically form federations. To perform FL, the framework deploys the needed services dynamically, monitors the training rounds, and calculates agents' trust scores based on the established commitments. ART4FL exploits a blockchain network to maintain the trust scores, and it provides those scores to negotiating agents' during the federations' formation phase.

With the digitalization and the ever more computerization of personal vehicles, new attack surfaces are introduced, challenging the security of the in-vehicle network. There is never such a thing as fully securing any computer system, nor learning all the methods of attack in order to prevent a break-in into a system. Instead, with sophisticated methods, we can focus on detecting and preventing attacks from being performed inside a system. The current state of the art of such methods, named intrusion detection systems (IDS), is divided into two main approaches. One approach makes its models very confident of detecting malicious activity, however only on activities that has been previously learned by this model. The second approach is very good at constructing models for detecting any type of malicious activity, even if never studied by the model before, but with less confidence. In this thesis, a new approach is suggested with a redesigned architecture for an intrusion detection system called Multi-mixed IDS. Where we take a middle ground between the two standardized approaches, trying to find a combination of both sides strengths and eliminating its weaknesses. This thesis aims to deliver a proof of concept for a new approach in the current state of the art in the CAN-bus security research field. This thesis also brings up some background knowledge about CAN and intrusion detection systems, discussing their strengths and weaknesses in further detail. Additionally, a brief overview from a handpick of research contributions from the field are discussed. Further, a simple architecture is suggested, three individual detection models are trained and combined to be tested against a CAN-bus dataset. Finally, the results are examined and evaluated. The results from the suggested approach shows somewhat poor results compared to other suggested algorithms within the field. However, it also shows some good potential, if better decision methods between the individual algorithms that constructs the model can be found. 

Termen Dark Patterns började användas år 2012 och har sedan dess blivit ett omdebatterat ämne. Termen kan översättas till manipulerande designval som medvetet utformas för att påverka och styra användarna av applikationer, hemsidor och sociala medier. Forskare menar på att det kan användas för att påverka människor till att ta val de annars aldrig gjort och flertalet granskningar av befintliga media visar på att deras närvaro är hög. Dark patterns fungerar som ett paraplybegrepp över dessa designmönster och har bryts ned av flera forskare för att kunna förklara varje enskilt identifierat mönster och dess påverkan. Tidigare forskning har dock inte fokuserat på Dark Patterns i samband med sociala medier med inriktning korta videoklipp, så kallade video shorts. Den här uppsatsen ämnar att ta reda på hur användarens upplevelse av Dark Patterns närvaro i denna kontext påverkar användaren. För att ta reda på detta används en designbaserad forskningsmetod. Genom intervjuer samlas kvalitativ data in från tio deltagare som delats upp i två grupper. En grupp har fått använda en prototyp med Dark Patterns som författarna till uppsatsen identifierat vid en analys av de sociala medierna TikTok och Youtube Shorts. Den andra prototypen innehåller designval som agerar motsats till dessa Dark Pattern. Uppsatsen använder sedan en tematisk analys för att kunna se likheter samt olikheter i deltagarnas svar.  I resultatet presenteras sedan vilka mönster analysen givit, exempelvis föredrar flertalet deltagare designen som innehåller Dark Patterns till den grad att de inte känner sig uppenbart lurade. Subtila Dark Patterns visar sig påverka användarens upplevelse positivt, vilket får dem att vilja använda tjänsten mer, vilket också ligger i tjänstens intresse. Det mer aggressiva och uppenbara Dark Pattern upprörde dock deltagarna i större utsträckning och kan bidra till att användaren känner sig lurad och känner sig istället frustrerad och irriterad. 

The rapidly evolving Internet of Things (IoT) systems demands addressing new requirements. This particularly needs efficient deployment of IoT systems to meet the quality requirements such as latency, energy consumption, privacy, and bandwidth utilization. The increasing availability of computational resources close to the edge has prompted the idea of using these for distributed computing and storage, known as edge computing. Edge computing may help and complement cloud computing to facilitate deployment of IoT systems and improve their quality. However, deciding where to deploy the various application components is not a straightforward task, and IoT system designer should be supported for the decision.

To support the designers, in this thesis we focused on the system qualities, and aimed for three main contributions. First, by reviewing the literature, we identified the relevant and most used qualities and metrics. Moreover, to analyse how computer simulation can be used as a supporting tool, we investigated the edge computing simulators, and in particular the metrics they provide for modeling and analyzing IoT systems in edge computing. Finally, we introduced a method to represent how multiple qualities can be considered in the decision. In particular, we considered distributing Deep Neural Network layers as a use case and raked the deployment options by measuring the relevant metrics via simulation.

Many Internet of Things (IoT) systems generate a massive amount of data needing to be processed and stored efficiently. Cloud computing solutions are often used to handle these tasks. However, the increasing availability of computational resources close to the edge has prompted the idea of using these for distributed computing and storage. Edge computing may help to improve IoT systems regarding important quality attributes like latency, energy consumption, privacy, and bandwidth utilization. However, deciding where to deploy the various application components is not a straightforward task. This is largely due to the trade-offs between the quality attributes relevant for the application. We have performed a systematic mapping study of 98 articles to investigate which quality attributes have been used in the literature for assessing IoT systems using edge computing. The analysis shows that time behavior and resource utilization are the most frequently used quality attributes; further, response time, turnaround time, and energy consumption are the most used metrics for quantifying these quality attributes. Moreover, simulation is the main tool used for the assessments, and the studied trade-offs are mainly between only two qualities. Finally, we identified a number of research gaps that need further study.

The deployment of Internet of Things (IoT) applications is complex since many quality characteristics should be taken into account, for example, performance, reliability, and security. In this study, we investigate to what extent the current edge computing simulators support the analysis of qualities that are relevant to IoT architects who are designing an IoT system. We first identify the quality characteristics and metrics that can be evaluated through simulation. Then, we study the available simulators in order to assess which of the identified qualities they support. The results show that while several simulation tools for edge computing have been proposed, they focus on a few qualities, such as time behavior and resource utilization. Most of the identified qualities are not considered and we suggest future directions for further investigation to provide appropriate support for IoT architects.

For the efficient execution of Deep Neural Networks (DNN) in the Internet of Things, computation tasks can be distributed and deployed on edge nodes. In contrast to deploying all computation to the cloud, the use of Distributed DNN (DDNN) often results in a reduced amount of data that is sent through the network and thus might increase the overall performance of the system. However, finding an appropriate deployment scenario is often a complex task and requires considering several criteria. In this paper, we introduce a multi-criteria decision-making method based on the Analytical Hierarchy Process for the comparison and selection of deployment alternatives. We use the RECAP simulation framework to model and simulate DDNN deployments on different scales to provide a comprehensive assessment of deployments to system designers. In a case study, we apply the method to a smart city scenario where different distributions and deployments of a DNN are analyzed and compared.

Although smart homes are tasked with an increasing number of everyday activities to keep users safe, healthy, and entertained, privacy concerns arise due to the large amount of personal data in flux. Privacy is widely acknowledged to be contextually dependent, however, the interrelated stakeholders involved in developing and delivering smart home services – IoT developers, companies, users, and lawmakers, to name a few – might approach the smart home context differently. This paper considers smart homes as digital ecosystems to support a contextual analysis of smart home privacy. A conceptual model and an ecosystem ontology are proposed through design science research methodology to systematize the analyses. Four privacy-oriented scenarios of surveillance in smart homes are discussed to demonstrate the utility of the digital ecosystem approach. The concerns pertain to power dynamics among users such as main users, smart home bystanders, parent-child dynamics, and intimate partner relationships and the responsibility of both companies and public organizations to ensure privacy and the ethical use of IoT devices over time. Continuous evaluation of the approach is encouraged to support the complex challenge of ensuring user privacy in smart homes.

This thesis explores the use of machine learning (ML) and smartphone sensors to improve indoordistance estimation, a critical aspect of healthcare tests like the 6-minute walk test (6MWT). In order to make tests like the 6MWT more available, and lower the barrier for a patient toget tested, there are multiple problems which need to be solved: How can the distance data needed for these tests be collected reliably and remotely, and without having to rely on the patient reporting correct data; How can these tests be performed indoors, without relying on GPS or other GNSS, which are unreliable indoors. To tackle these challenges a convolutional neural network (CNN) trained on a dataset containing continuous ground truth was employed. An enhancement of an existing CNN model was done by collecting more training data, tuning hyper parameters, and testing it on a diverse dataset.

The results of this thesis shows that when predicting distance walked on data from participants the CNN model has seen before, the precision meets clinical minimum for being able to show a change in the health condition. On real world data the performance suffers. Despite limitations due to the scope of data collection, the results still underscores the potential of ML for accurate and efficient indoor distance estimation and points to future research directions.

This structured literature study examines which tools and technologies related to artificialintelligence (AI) that can support teachers by replacing repetitive and time-consumingadministrative tasks that are outside of the direct interaction between teacher and student.

By reviewing available research in this field, we map these findings. This is then followedby a discussion of the results and potential of the latest technology, as well as an indicationof where further research can take place.To map existing research on the subject, this literature study has been conducted basedon a general course divided into three phases: before, during, and after. Due to the po-tential size of this area, we investigated only one case per each phase, respektively coursedesign, grading and course evaluation. The literature study is limited to higher educationinstitutions such as universities and colleges.

Our results shows that the majority of AI techniques and tools used in these three areasare based on natural language processing (NLP) and neural networks. We found that thereis a greater amount of research focused on grading compared to course design and courseevaluation, and that research in all these areas is increasing over time. However, we note alack of research investigating the practical usage of these tools and techniques by teachers.

Designing entertaining games requires control over the difficulty experienced by players. Gameplay has a clear impact on the difficulty and as such is easy to handle but other factors such as graphics do not share this clear cut impact. Therefore a game was made with adjustable graphical settings without changing any game elements in order to evaluate the effect on 18 participants. No effect was found for the measured performance but a weak correlation is possible for perceived difficulty.

The smart home technology market is witnessing rapid growth due to the advent of more advanced, intuitive, and affordable solutions. As the adoption of these technologies becomes more prevalent, there is a need for research to explore potential avenues for pervasive smart living. This study aims to review the available literature and industry studies, along with our own experiences in the field, to identify and discuss potential future research in the smart home. We observe that the future of the smart home will likely be focused on improving the user experience, with a greater emphasis on personalization, automation, and Artificial intelligence (AI)-driven technologies, leading to what we call the "Green Intelligent Home". Through this analysis, this study aims to offer insights into how the development of smart homes could shape society in the future and the potential implications of such a development. This study concludes by suggesting a framework for knowledge development in the smart home domain.

Smart homes have become increasingly popular for IoT products and services with a lot of promises for improving the quality of life of individuals. Nevertheless, the heterogeneous, dynamic, and Internet-connected nature of this environment adds new concerns as private data becomes accessible, often without the householders’ awareness. This accessibility alongside with the rising risks of data security and privacy breaches, makes smart home security a critical topic that deserves scrutiny. In this paper, we present an overview of the privacy and security challenges directed towards the smart home domain. We also identify constraints, evaluate solutions, and discuss a number of challenges and research issues where further investigation is required.

The Internet of Things (IoT) unleashes great opportunities to improve our way of living and working through a seamless and highly dynamic cooperation among heterogeneous things including both computer-based systems and physical objects. However, properly dealing with the design, development, deployment and runtime management of IoT applications means to provide solutions for a multitude of challenges related to intelligent distributed systems within the IoT. In this paper we propose Model-Driven Engineering (MDE) as a key enabler for applications running on intelligent distributed loT systems. MDE helps in tackling challenges and supporting the lifecycle of such systems. Specifically, we introduce MDE4IoT, an MDE approach enabling the modelling of things and supporting intelligence as self-adaptation of Emergent Configurations in the IoT. Moreover, we show how MDE, and in particular MDE4IoT, can help in tackling several challenges by providing the Smart Street Lights concrete case.

The researchers in the DESIRE4EU project propose the use of an eco-optimized PLA/Flax-based PCB manufacturing substrate that could be recycled or degraded after use due to the organic, biodegradable nature of the material. However, having a biodegradable PCB is not enough, as we need to rethink electronics design to better fit specific use cases, avoiding a one-size-fits-all philosophy. This paper includes a set of learned lessons in the form of design guidelines extracted from the experience of a small-scale manufacturing of a set of biodegradable microcontroller boards. These lessons outline contemporary technical limitations of biodegradable PCBs, which the authors trust will be overcome during the development of the mentioned EU project by 2027. This should be a first step towards reducing e-waste in the not-so-far future. This positioning paper states that, in the current ecological crisis, the different engineering communities need to rethink their priorities in order to produce ecology-friendly innovations by keeping concepts such as just enough computing and ecooptimization in mind. In particular, when talking about the TinyML community, we suggest carefully considering the hardware’s limiting factors presented by computational power, or radio communication when designing new Edge devices so that they could use 2-layers biodegradable PCBs. In parallel to the hardware discussion, the authors bring up issues emerging from using bloated inference software production workflows, which have a very direct impact in the ecology due to the computation power needed for embedded machine learning software production. This paper suggests the adoption of ondevice training to minimize the energy consumption and dependance on connected toolchains during programming. 

Digitalization is a term that is frequently mentioned in discussions about efficiency improvement opportunities. In order to digitalize, competence is required, which leads to a greater need for competence and thus a stronger focus on competence supply. Digitalization is central when technology is used to improve and improve the efficiency of operations. In order to develop and implement digital solutions, it is crucial to have sufficient skills and knowledge within the organization. The research problem is of great importance because it addresses the challenges that public organizations face in terms of competence supply in combination with digitalization. The success of digitalization requires that organizations have sufficient competence in digital tools and technologies. The challenges posed by competence supply can be crucial to the success of digitalization projects and initiatives. Using a qualitative research strategy and semi-structured interviews, the challenges faced by the region in the digitalization work have been studied, mainly in relation to competence supply. In the collected data material, it was possible to note three major challenges experienced by the region: lack of expertise, attitudes towards changes within the organization and challenges in the implementation of projects. The study also shows that the willingness to digitalize is high but that they have not yet managed to maximize the value of digitalizing analogue processes - thus there is room for improvement in the region in terms of maximizing the potential of digitalization.

This study presents lessons learned based on practical experiences of connecting devices to internet-of-things platforms in the context of research and academic coursework. The experiences are gathered from six research projects, one undergraduate course, and a few undergraduate theses over a three-year period. The lessons learned include: the trade-off of rapid prototyping over security is very common, example source code is not up to production standards, adherence to standards speeds development, debugging support for IoT systems is lacking, open source licenses varies, poor platform interoperability, and the array of service fees among platform providers obstruct cost comparisons.

As urbanization and densification increases there appears to be a reduction in available green space. This is a concern as links between people's health and exposure to greenspaces have been found. The perceived sensory dimensions (PSDs), which consist of eight environmental qualities that support people's needs in various ways, have been proposed as a way of helping in the designing and planning of urban green spaces. In this study, we explore the feasibility of training a generative model to generate images based on values representing the PSDs. We hope that this could serve as a possible tool to help in designing greenspaces in the future. To evaluate our model we make use of a questionnaire. The results of the questionnaire suggest that combining the PSDs with a generative model is likely to be possible and something that could be worth exploring further.

The North Atlantic Right Whale (NARW) population is currently teetering on the brink of extinction, with a mere approximate count of 350 individuals remaining. These animals have been protected under the Endangered Species Act since 1970. Today, the survival of right whales is imperiled primarily due to vessel collisions, net entanglements, and habitat degradation. This paper presents a novel system of animal-computer interaction founded on the identification of bioacoustic signatures. Initially, NARWs' vocalizations were transformed into spectrograms, which were subsequently inputted into a Convolutional Neural Network (CNN). To enhance robustness against environmental noise, techniques such as time warping, frequency masking, and time masking were employed. The outcomes of our study indicate that the proposed system holds potential for establishing a closed-loop interaction framework between vessels and NARWs. This framework could enable vessels to adapt their speed or avoid routes frequented by NARWs. Furthermore, this article discusses the potential benefits of employing networked sensors, such as Internet of Things (IoT) devices, to augment NARW monitoring and data collection efforts.

Timed up and go (TUG) test is a widely recognised and standardised mobility test to measure basic mobility and balance capabilities. Although its reliability is proven, the acquisition of ac- curate and objective information requires technical equipment. To address this need for accuracy and objectivity, our research examines the utilisation of non-wearable sensors for gait analysis during a TUG test. Collaborating with the Italian National Research Council (CNR) in Pisa, the study illustrates our attempt to develop software specifically designed for conducting the TUG test in conjunction with hardware from SensingTex’s Sensing Mat portfolio. By doing this, the re- search also aims to fill the gap in TUG-specific software targeting sensorised mats. The developed software characterise the gait during the TUG test by identifying and examining the following parameters: TUG-time, Sit-to-Stand, Mid-Turning, and End-Turning-Stand-to-Sit. Additionally, meaningful parameters for overall gait assessment have been selected: walking speed and stride length. Two experimental iterations were conducted to assess the reliability of the developed soft- ware. Both iterations involved two different groups of six healthy participants (41.58±13.32 yrs; 6 females, 6 males) performing various walking types. The results indicates that by translating these observations into quantitative data, our research has the potential to enhance the accuracy and objectivity of gait analysis, thereby improve clinical evaluations and advancing the field.

VPN is primarily used to encrypt your network traffic and identity online securely from a private location. This can be used as a safety measure to prevent theft of personal data. It also allows its user to change the geolocation to wherever they want which unlocks the possibility to use another country's services. Related work has shown that there are also downsides to using VPN services. Some VPN solutions do have security problems that its user could be unaware of. This study explored the experience and beliefs surrounding the usage of VPN while browsing the internet from people with software expertise. Interviews were conducted with people in different areas surrounding usage of VPN services to get a deeper understanding of why VPN is used and to what extent they believe VPN is providing anonymity and security of their data. The findings from this study is that the main reason to use a VPN is to access unavailable services. These services can vary from content online that is not available in the region from where you access the internet to services that are work related and locked to specific networks. Another finding was also that among these people the belief that the use of a VPN was enough to make a user anonymous by itself is controversial.

Since the advent of Artificial Intelligence (AI) and Machine Learning (ML), researchers have asked how intelligent computing systems could interact with and relate to their users and their surroundings, leading to debates around issues of biased AI systems, ML black-box, user trust, user’s perception of control over the system, and sys- tem’s transparency, to name a few. All of these issues are related to how humans interact with AI or ML systems, through an interface which uses different interaction modalities. Prior studies address these issues from a variety of perspectives, spanning from under- standing and framing the problems through ethics and Science and Technology Studies (STS) perspectives to finding effective technical solutions to the problems. But what is shared among almost all those efforts is an assumption that if systems can explain the how and why of their predictions, people will have a better perception of control and therefore will trust such systems more, and even can correct their shortcomings. This research field has been called Explainable AI (XAI). In this studio, we take stock on prior efforts in this area; however, we focus on using Tangible and Embodied Interaction (TEI) as an interaction modality for understanding ML. We note that the affordances of physical forms and their behaviors potentially can not only contribute to the explainability of ML sys- tems, but also can contribute to an open environment for criticism. This studio seeks to both critique explainable ML terminology and to map the opportunities that TEI can offer to the HCI for designing more sustainable, graspable and just intelligent systems.

Machine Learning (ML) has been a prominent area of research within Artificial Intelligence (AI). ML uses mathematical models to recognize patterns in large and complex data sets to aid decision making in different application areas, such as image and speech recognition, consumer recommendations, fraud detection and more. ML systems typically go through a training period in which the system encounters and learns about the data; further, this training often requires some degree of human intervention. Interactive machine learning (IML) refers to ML applications that depend on continuous user interaction. From an HCI perspective, how humans interact with and experience ML models in training is the main focus of this workshop proposal. In this workshop we focus on the user experience (UX) of Interactive Machine Learning, a topic with implications not only for usability but also for the long-term success of the IML systems themselves.

Peer-to-peer (P2P) reputation systems, such as those used by eBay and Amazon, servean important role on the web, especially in E-commerce, as online reputation serves asa primary guiding factor for consumers in making informed decisions. The importanceof these systems, and also the increasing popularity of P2P and distributed systems, theissue of how to prevent and resist sybil and re-entry attacks becomes an important area ofresearch as they can impinge the integrity of those systems. To address this issue, in thisthesis, we propose an approach that encompasses a software architecture and processeswhich serves as a proof-of-concept of how to mitigate sybil and re-entry attacks on review based P2P distributed reputation systems. The architecture uses novel technologiessuch as blockchain, smart contracts, and non-fungible tokens (NFT) in conjunction withSwedish E-id provider BankID to build a sybil and re-entry attack resistant reputationsystem. To validate the feasibility of our approach, we developed a prototype and used itto run experiments to evaluate the functional correctness of the architecture as a mitigation solution

This paper is aimed at investigating whether a biofeedback device can positively increase a player's experience of thrill and suspension in a horror game. To facilitate this, two versions of the same horror game were created with a connection to a heart rate monitor. The difference between the two versions of the game was that the core elements were controlled by the heart rate which attempted to keep the player in a constant suspense state based on their heart rate. This was done to enhance the player's experience and thrill. These two versions of the game were then play-tested by users. Users had no insight into which version they were testing and afterward, a questionnaire was administered to ascertain the tester's emotional responses.The collected data was then analyzed and a pattern could be observed where testers preferred the version of the game that was controlled by the heart rate. This, backed by previous studies showed that using a biofeedback device to implement only the heart rate into a game had a marked positive effect on player experience. Still, additional research is needed with a larger control group to get more accurate results.

Mobile apps are an increasingly important part of public transport, and can be seen as part of the journey experience. Personalisation of the app is then one aspect of the experience that, for example, can give travellers a possibility to save favourite journeys for easy access. Such a list of journeys can be extensive and inaccurate if it doesn’t consider the traveller’s context. Making an app context aware and present upcoming journeys transforms the app experience in a personal direction, especially for commuters. By using historical personal contextual data, a travel app can present probable journeys or accurately predict and present an upcoming journey with departure times. The predictions can take place when the app is started or be used to remind a commuter when it is time to leave in order to catch a regularly travelled bus or train.

To address this research opportunity we have created an individually trained Machine Learning (ML) agent that we added to a publicly available commuter app. The added part of the app uses weekday, time, user activity and location to predict a user’s upcoming journey. Predictions are made when the app starts and departure times for the most probable transport are presented to the traveller. In our case a commuter only makes a few journey searches in the app every day which implies that, based on our contextual parameters, it will take at least some weeks to create journey patterns that can give acceptable accuracy for the predictions. In the work we present here, we focus on how to handle this cold start problem e.g. the situation when no or inaccurate historical data is available for the Machine Learning agent to train from. These situations will occur both initially when no data exists and due to concept drift originating from changes in travel patterns. In these situations, no predictions or only inaccurate predictions of upcoming journeys can be made.    

We present experiences and evaluate results gathered when designing the interactions needed for the MT session as well as design decisions for the ML pipeline and the ML agent. The user’s interaction with the ML agent during the teaching session is a crucial factor for the success. During the teaching session, information on what the agent already has learnt has to be presented to the user as well as possibilities to unlearn obsolete commute patterns and to teach new. We present a baseline that shows an idealised situation and the amount of training data that the user needs to add in a MT session to reach acceptable accuracy in predictions. Our main contribution is user evaluated design proposals for the MT session.

Using individually trained ML agents opens up opportunities to protect personal data and this approach can be used to create mobile applications that is independent of local transport providers and thus act on open data on a global scale.

Contemporary Machine Learning (ML) often focuseson large existing and labeled datasets and metrics aroundaccuracy and performance. In pervasive online systems, conditionschange constantly and there is a need for systems thatcan adapt. In Machine Teaching (MT) a human domain expertis responsible for the knowledge transfer and can thus addressthis. In my work, I focus on domain experts and the importanceof, for the ML system, available features and the space they span.This space confines the, to the ML systems, observable fragmentof the physical world. My investigation of the feature space isgrounded in a conducted study and related theories. The resultof this work is applicable when designing systems where domainexperts have a key role as teachers.

Machine learning research today is dominated by atechnocentric perspective and in many cases disconnected fromthe users of the technology. The machine teaching paradigm insteadshifts the focus from machine learning experts towards thedomain experts and users of machine learning technology. Thisshift opens up for new perspectives on the current use of machinelearning as well as new usage areas to explore. In this study,we apply and map existing machine teaching principles ontoa contextual machine teaching implementation in a commutingsetting. The aim is to highlight areas in machine teaching theorythat requires more attention. The main contribution of this workis an increased focus on available features, the features space andthe potential to transfer some of the domain expert’s explanatorypowers to the machine learning system.

This paper addresses the problem of hazard recognition and risk assessment in open and non-predictive environments to support decision making and action selection. Decision making and action selection incorporate decreasing situational risks and maintain safety as operational constraints. Commonly, neither existing application-related safety standards nor the situation modeling or knowledge representation is considered in that context. This contribution introduces a novel approach denoted as a Safety-Driven Behavior Management focusing on situation modeling and the problem of knowledge representation in its sub-functions in the context of situational risks. It combines the safety standards-oriented hazard analysis and the risk assessment approach with the machine learning-based situation recognition. An example illustrating the approach is presented in this paper.

The ownership and demand for fitness trackers, smartwatches, and wrist-worn deviceshave been increasing globally. These devices offer various features such as measuringphysical activity, sleep monitoring, and health-related measurements like heart rate andheart rate variability using PhotoPlethysmoGraphy (PPG). However, research indicatesthat PPG measurements are less accurate on darker skin compared to lighter skin due to thehigher presence of melanin, a light-absorbing substance in dark skin. This thesis addresses the impact of melanin on the accuracy of heart rate measurements on different skin colors using four commercial smartwatches. The study involves analyzing the accuracy of these smartwatches on individuals with varying skin colors while controlling for external factors. The collected data from the smartwatches are compared to a reference sensor that uses electrocardiography (ECG) measurements with electrodes placed around the chest. Three different tests are conducted wearing the devices, with no movement, while walking, and with circular hand motions.The tests were conducted on twelve participants representing the 6 different skin types categorized using the Fitzpatrick scale. With the presented results in this thesis, it was concluded that the 4 smartwatches' measurement accuracy does not seem to be dependent on specific skin types. 

This Bachelor’s thesis provides a comparative analysis of the core algorithms of A-star (A*) and Breadth-First Search (BFS) in multi-target scenarios. Previous research has conducted comparisons of these algorithms in single-target scenarios and improvementst o the algorithms to address their limitations. However, this thesis evaluates the basic versions of A* and BFS for situations where complex implementations are not possible or preferred. The study systematically simulates these algorithms across various scenarios to understand their performance in managing multiple targets. The results show that BFS is generally more effective in scenarios with a small search space and in environments with a higher number of targets due to its ability to locate multiple targets in a single search. Conversely, A* performs better in scenarios where there are fewer targets and the search space is larger, due to its heuristic approach that prioritizes paths which seem more promising. The thesis provides guidelines for developers and researchers to assist in the decision-making process when choosing between these two algorithms depending on specific application requirements.

Denna studies syfte är att ställa Progressive web apps mot native applikationer för att jämföra upplevd user experience.

Progressive Web App är en teknik som börjar bli allt mer vanlig för företag att implementera. Författarna utforskar ämnet mer riktat åt användarens upplevelse för att se vilka faktorer som kan spela in i valet mellan Progressive web app och native.

Författarna har använt sig flertalet olika metoder för att studera om det finns någon skillnad i användarupplevelsen mellan en Progressive web app och en native applikation. Dessa metoder innefattar exempelvis think aloud protocol, system usability scale och Nielsens heuristics. Sett till teorin bakom en Progressive web app så säger den att det ska vara möjligt att kunna replikera en native applikation i form av funktionalitet men vi har i vår studie valt att titta utanför funktionalitet och lagt mer fokus på användarupplevelsen.

Studien påvisar att de finns skillnader i upplevd användbarhet mellan PWA och native i vissa avseenden. Men de går inte att helt fastställa om dessa beror på utomstående faktorer.

Trots att studien inte påvisade ett klart svar så ger de en grund till vidare forskning för hur man kan välja mellan de olika teknikerna framåt.

With the increasing diffusion of Internet of Things (IoT) technologies, the transportation of goods sector is in a position to adopt novel intelligent services that cut across the otherwise highly fragmented and heterogeneous market, which today consists of a myriad of actors. Legacy systems that rely upon direct integration between all actors involved in the transportation ecosystem face considerable challenges for information sharing. Meanwhile, IoT based services, which are designed as devices that follow goods and communicate directly to cloud-based backend systems, may provide services that previously were not available. For the purposes of this paper, we present a theoretical framework for classification of such intelligent goods systems based on a literature study. The framework, labelled as the Intelligent Goods Service (IGS) framework, aims at increasing the understanding of the actors, agents, and services involved in an intelligent goods system, and to facilitate system comparisons and the development of new innovative solutions. As an illustration of how the IGS framework can be used and contribute to research in this area, we provide an example from a direct industry-academia collaboration.

Context: With the advent of Machine Learning (ML) and especially Deep Learning (DL) technology, companies are increasingly using Artificial Intelligence (AI) in systems, along with electronics and software. Nevertheless, the end-to-end process of developing, deploying and evolving ML and DL models in companies brings some challenges related to the design and scaling of these models. For example, access to and availability of data is often challenging, and activities such as collecting, cleaning, preprocessing, and storing data, as well as training, deploying and monitoring the model(s) are complex. Regardless of the level of expertise and/or access to data scientists, companies in all embedded systems domain struggle to build high-performing models due to a lack of established and systematic design methods and processes.

Objective: The overall objective is to establish systematic and structured design methods and processes for the end-to-end process of developing, deploying and successfully evolving ML/DL models.

Method: To achieve the objective, we conducted our research in close collaboration with companies in the embedded systems domain using different empirical research methods such as case study, action research and literature review.

Results and Conclusions: This research provides six main results: First, it identifies the activities that companies undertake in parallel to develop, deploy and evolve ML/DL models, and the challenges associated with them. Second, it presents a conceptual framework for the continuous delivery of ML/DL models to accelerate AI-driven business in companies. Third, it presents a framework based on current literature to accelerate the end-to-end deployment process and advance knowledge on how to integrate, deploy and operationalize ML/DL models. Fourth, it develops a generic framework with five architectural alternatives for deploying ML/DL models at the edge. These architectural alternatives range from a centralized architecture that prioritizes (re)training in the cloud to a decentralized architecture that prioritizes (re)training at the edge. Fifth, it identifies key factors to help companies decide which architecture to choose for deploying ML/DL models. Finally, it explores how MLOps, as a practice that brings together data scientist teams and operations, ensures the continuous delivery and evolution of models. 

Companies across domains are rapidly engaged in shifting computational power and intelligence from centralized cloud to fully decentralized edges to maximize value delivery, strengthen security and reduce latency. However, most companies have only recently started pursuing this opportunity and are therefore at the early stage of the cloud-to-edge transition. To provide an overview of AI deployment in the context of edge/cloud/hybrid architectures, we conduct a systematic literature review and a grey literature review. To advance understanding of how to integrate, deploy, operationalize and evolve AI models, we derive a framework from existing literature to accelerate the end-to-end deployment process. The framework is organized into five phases: Design, Integration, Deployment, Operation and Evolution. We make an attempt to analyze the extracted results by comparing and contrasting them to derive insights. The contribution of the paper is threefold. First, we conduct a systematic literature review in which we review the contemporary scientific literature and provide a detailed overview of the state-of-the-art of AI deployment. Second, we review the grey literature and present the state-of-practice and experience of practitioners while deploying AI models. Third, we present a framework derived from existing literature for the end-to-end deployment process and attempt to compare and contrast SLR and GLR results.

Multi-task learning has become a powerful solution in which multiple tasks are trained together to leverage the knowledge learned from one task to improve the performance of the other tasks. However, the tasks are not always constructive on each other in the multi-task formulation and might play negatively during the training process leading to poor results. Thus, this study focuses on finding the optimal group of tasks that should be trained together for multi-task learning in an automotive context. We proposed a multi-task learning approach to model multiple vehicle long-term behaviors using low-resolution data and utilized gradient descent to efficiently discover the optimal group of tasks/vehicle behaviors that can increase the performance of the predictive models in a single training process. In this study, we also quantified the contribution of individual tasks in their groups and to the other groups' performance. The experimental evaluation of the data collected from thousands of heavy-duty trucks shows that the proposed approach is promising.

N/A.

This book contains the proceedings of the 9th International Conference on the Internet of Things, Big Data and Security (IoTBDS 2024). This year, IoTBDS was held in Angers, France, from April 28 - 30, 2024. It was sponsored by the Institute for Systems and Technologies of Information, Control and Communication (INSTICC). The Internet of Things (IoT) is a platform that allows a network of devices (sensors, smart meters, etc.) to communicate, analyze data and process information collaboratively in the service of individuals or organizations'. The IoT network can generate large amounts of data in a variety of formats and use different protocols, which can be stored and processed in the cloud. The conference looks to address the issues surrounding IoT devices, their interconnectedness and the services they may offer, including efficient, effective and secure analysis of the data IoT produces using machine learning and other advanced techniques, models and tools, and issues of security, privacy and trust that will emerge as IoT technologies mature and become part of our everyday lives. Big Data (BD) has core values of volume, velocity, variety and veracity. After collecting much data from IoT, BD can be jointly used with machine learning, AI, statistical and other advanced techniques, models and methods, which can create value for people and organizations adopting it, since forecasting, deep analysis and analytics can help identify weaknesses and make improvements based on different analysis. Maintaining a high level of security and privacy for data in IoT is crucial, and we welcome recommendations, solutions, demonstrations and best practices for all forms of security and privacy for IoT and BD. IoTBDS 2024 received 51 paper submissions from 22 countries of which 20% were accepted and published as full papers. A double-blind paper review was performed for each submission by at least 2 but usually 3 or more members of the International Program Committee, which is composed of established researchers and domain experts. The high quality of the IoTBDS 2024 program is enhanced by the keynote lecture delivered by distinguished speakers who are renowned experts in their fields: Luigi Atzori (Università degli Studi di Cagliari, Italy), Patrick Hung (Faculty of Business and IT, Ontario Tech University, Canada), Matthieu Deboeuf Rouchon (Capgemini Engineering, France) and Samuel Fosso Wamba (Toulouse Business School, France). The conference is complemented by a Workshop on Collaborative EU Research Projects, chaired by Victor Chang and Jia-Chun Lin. All presented papers will be available at the SCITEPRESS Digital Library and will be submitted for evaluation for indexing by SCOPUS, Google Scholar, The DBLP Computer Science Bibliography, Semantic Scholar, Engineering Index and Web of Science / Conference Proceedings Citation Index. As recognition for the best contributions, several awards based on the combined marks of paper reviewing, as assessed by the Program Committee, and the quality of the presentation, as assessed by session chairs at the conference venue, are conferred at the closing session of the conference. A shortlist of papers presented at the conference will be selected for publication of extended and revised versions in the special issues of the Springer Nature Computer Science Journal, Journal of Global Information Management, Big Data Journal and Internet of Things. The program for this conference required the dedicated effort of many people. Firstly, we must thank the authors, whose research efforts are herewith recorded. Next, we thank the members of the Program Committee and the auxiliary reviewers for their diligent and professional reviewing. We would also like to deeply thank the invited speakers for their invaluable contribution and for taking the time to prepare their talks. Finally, a word of appreciation for the hard work of the INSTICC team; organizing a conference of this level is a task that can only be achieved by the collaborative effort of a dedicated and highly capable team. We wish you all an exciting and inspiring conference. We hope to have contributed to the development of our research community, and we look forward to having additional research results presented at the next edition of IoTBDS, details of which are available at https://iotbds.scitevents.org.

We present a prototype of a system for machine learning (ML) powered interactive generative literature called Multiverse. The system employs a set of neural networks models to dynamically generate a literary space from an initial writing prompt provided by its user-reader. The user-reader is able to choose the model used to generate the text as a kind of interactive machine learning (IML). The research explores how interaction design and HCI researchers can engage directly with ML by leveraging the powerful, yet accessible, models afforded by new developments in the field. User-readers testing the prototype found the imperfect aesthetics of the ML-generated texts to be entertaining and engaging but struggled to conceptualize the generated work as a navigable interactive literary space.

Text compression schemes and succinct data structures usually combine sophisticated probability modes with basic coding methods whose average codeword length closely match the entropy of known distributions. In the frequent case where basic coding represents run-lengths of outcomes with probability p, i.e. geometric distribution Pr(i)=pⁱ(1-p), a Golomb code is an optimal instantaneous code, which has the additional advantage that codewords can be computed using only an integer parameter calculated from p, without need for a large or sophisticated data structure. Golomb coding does not, however, gracefully handle the case where run-lengths are bounded by a known integer n, where codewords allocated for the case i>n are wasted. While negligible for large n, this makes Golomb coding unattractive in situations where n is recurrently small, e.g., when representing many short lists, or when the range of n is narrowed down by a recursive algorithm.

With the growing use of AI tools in the education sector, we are faced with a crucial question: should AI tools be allowed and integrated into the Swedish education system? The purpose of this study is to examine and analyze how programming students in the software development program use AI tools to solve programming tasks, as well as to map out the perceived benefits and challenges experienced by students. The study also compares specialized programming tools, such as Github Copilot and Tabnine, with general AI tools like ChatGPT. 

The study employed a research methodology involving an empirical investigation through a qualitative analysis of interviews with second- and third-year students in the software development program, as well as a quantitative survey of students who were assigned an AI tool in the courses DA489A System Development II and DA109A Web Services at Malmö University. This was supplemented by an extensive literature overview.The aim of these methodologies was to achieve a deeper understanding of students' experiences with AI tools in their studies, as well as to provide a broader perspective through generalizations based on numerical data.

The results show that AI-tools have a positive impact on the educational sector. Students report a increase in productivity, efficiency, learning, higher coding speeds, and motivation. AI-tools are primarily used to generate simpler code or code skeletons, which saves time for students to focus more on the more complex tasks. The tools are also used for debugging, syntax assistance, translation between programming languages, idea development, information retrieval, as a programming partner (Copilot), and as a 'rubber duck.'

The conclusion is that AI tools have the potential to positively impact the education, but careful consideration and guidelines are necessary to ensure their ethical and effective use. This study contributes to the ongoing debate on digitalization in education and provides insights that may be valuable to policymakers in the education sector.

Adoption of electric vehicles has been restrained by the availability of charging stations and consumer fear of being stranded with a depleted battery, far from the nearest charger. In many areas of the world, charging stations are now widely available and the transition from vehicles with internal combustion engines is accelerating, though still in a fairly early stage. For electric vehicle drivers in those areas, anxiety that they will not be able to find a charger (“range anxiety”) is subsiding. However, differences in charging speed and pricing between stations and different outlets at the same station can be large. Total trip duration can vary significantly based on the charging outlet selected. Prior research has developed methods for helping all drivers find the fastest route and for electric vehicle drivers to ensure that they are able to complete their trip. Additional research has explored other complexities of route selection for electric vehicles such as how to select optimal stations for charging based on the total trip duration, including driving and charging time. Pricing for recharging electric vehicles at public chargers is more complex and diverse than for gas filling stations due to the differences in charging rates and the relatively low competition. This research investigates those differences. Using design science research methodology, a method is presented for determining which charging stops result in the lowest possible charging cost for a given route. The method is demonstrated through experiment with random routes within Sweden. The experimental results show that the average cost savings as compared to the duration-optimal route is 15% and 139 SEK per additional hour of trip time. One possible direction for future work is to improve the performance of the algorithm for use in real-time consumer route planning applications.

The papers in this special section examines the deployment of Big Data and artificial intelligence for network technologies. The eneration of huge amounts of data, called big data, is creating the need for efficient tools to manage those data. Artificial intelligence (AI) has become the powerful tool in dealing with big data with recent breakthroughs at multiple fronts in machine learning, including deep learning. Meanwhile, information networks are becoming larger and more complicated, generating a huge amount of runtime statistics data such as traffic load, resource usages. The emerging big data and AI technologies may include a bunch of new requirements, applications and scenarios such as e-health, Intelligent Transportation Systems (ITS), Industrial Internet of Things (IIoT), and smart cities in the term of computing networks. The big data and AI driven network technologies also provide an unprecedented patient to discover new features, to characterize user demands and system capabilities in network resource assignment, security and privacy, system architecture, modeling and applications, which needs more explorations. The focus of this special section is to address the big data and artificial intelligence for network technologies. We appreciate contributions to this special section and the valuable and extensive efforts of the reviewers. The topics of this special section range from big data and AI algorithms, models, architecture for networks and systems to network architecture.

This thesis investigates the automation of forensic analysis in identifying and categorising forensically interesting changes across different versions of chat applications on Android platforms. The focus is primarily on the differentiation of Android Package Kit (APK) using reverse-engineering techniques to reconstruct the original source code and comparing the source code from two different versions of the APK. Given the rapid evolutionof chat applications and their frequent updates, it is crucial for forensic investigators to understand these changes to maintain the integrity of legal investigations.

The research introduces a comprehensive framework leveraging the open-source tools, Ghidra and BinDiff, to automate the decompilation and differential analysis of APK files. This approach not only makes forensic analysis less complicated but also ensures that investigators can keep pace with the continuous updates in chat applications.

Tests on the system are conducted on various versions of the Signal chat application. These tests aim to demonstrate the proposed tool in capturing significant changes between APK versions, such as alterations inlogging mechanisms, database interactions, and the use of encryption and cypher libraries.

The results confirm that the integration of Ghidra and BinDiff provides a solution for automated forensic analysis, facilitating the identification of changes and categorisation of methods based on their forensic relevance. The study shows that the tool can pinpoint modifications and structural changes, which are essential for forensic investigations.

Modern indoor positioning systems encounter challenges inherent to indoor environments. Signal changes can stem from various factors like object movement, signal propagation, or obstructed line of sight. This thesis explores a supervised machine learning approach that integrates Bluetooth Low Energy (BLE) and inertial sensor data to achieve consistent angle and distance estimations.

The method relies on BLE angle estimations and signal strength alongside additional sensor data from an Inertial Measurement Unit (IMU). Relevant features are extracted and a supervised learning model is trained and then validated on familiar environment tests. The model is then gradually introduced to more unfamiliar test environments, and its performance is evaluated and compared accordingly.

This thesis project was conducted at the u-blox office and presents a comprehensive methodology utilizing their existing hardware. Several extensive experiments were conducted, refining both data collection procedures and experimental setups. This iterative approach facilitated the improvement of the supervised learning model, resulting in a proposed model architecture based on transformers and convolutional layers. The provided methodology encompasses the entire process, from data collection to the evaluation of the proposed supervised learning model, enabling direct comparisons with existing angle estimation solutions employed at u-blox. The results of these comparisons demonstrate more accurate outcomes compared to existing solutions when validated in familiar environments. However, performance gradually declines when introduced to a new environment, encountering a wider range of signal conditions than the supervised model had trained on. Distance estimations are then compared with the path loss propagation equation, showing an overall improvement.