Roadside Unit-based Unknown Object Detection in Adverse Weather Conditions for Smart Internet of Vehicles

2023 | ACM Transactions on Management Information Systems | Citations: 0

Authors: Chen, Yu-Chia; Jhong, Sin-Ye; Hsia, Chih-Hsien

Abstract: For Internet of Vehicles applications, reliable autonomous driving systems usual ... Expand

Abstract: For Internet of Vehicles applications, reliable autonomous driving systems usually perform the majority of their computations on the cloud due to the limited computing power of edge devices. The communication delay between cloud platforms and edge devices, however, can cause dangerous consequences, particularly for latency-sensitive object detection tasks. Object detection tasks are also vulnerable to significantly degraded model performance caused by unknown objects, which creates unsafe driving conditions. To address these problems, this study develops an orchestrated system that allows real-time object detection and incrementally learns unknown objects in a complex and dynamic environment. A you-only-look-once–based object detection model in edge computing mode uses thermal images to detect objects accurately in poor lighting conditions. In addition, an attention mechanism improves the system’s performance without significantly increasing model complexity. An unknown object detector automatically classifies and labels unknown objects without direct supervision on edge devices, while a roadside unit (RSU)-based mechanism is developed to update classes and ensure a secure driving experience for autonomous vehicles. Moreover, the interactions between edge devices, RSU servers, and the cloud are designed to allow efficient collaboration. The experimental results indicate that the proposed system learns uncategorized objects dynamically and detects instances accurately. Collapse

Topics: cloud computing autonomous driving system database system internet technology internet of things

Methods: computational algorithm cluster analysis design artifact experiment artificial neural network

Situational Factor Determinants of the Allocation of Decision Rights to Edge Computers

2023 | ACM Transactions on Management Information Systems | Citations: 0

Authors: Chua, Cecil Eng Huang; Niederman, Fred

Abstract: Internet of Things (IoT) designers frequently must determine whether action-orie ... Expand

Abstract: Internet of Things (IoT) designers frequently must determine whether action-oriented decisions should be made by edge computers or whether they should be made only by central servers combining input from all edge computers. An important example of this design problem occurs in fire protection IoT, where individual edge computers attached to sensors might be empowered to make decisions (have decision rights) about how to manage the fire. Alternatively, decision rights could be held exclusively by a central server isolated from the fire, because the designer is concerned damage to edge computers could cause them to act unreliably. This research models this allocation of decision rights to identify the relative influence of various decision factors. We first model the allocation of decision rights under the following assumptions: (1) The central server cannot make an error the edge computer cannot make; (2) the central server cannot update the edge computer with its information in a timely manner; and (3) the central server cannot reverse an action initiated by the edge computer to explore the factors impacting decision rights conferral. We then relax each of these three assumptions. We show how relaxing each assumption radically changes the factors impacting decision rights conferral. We also show that allowing the central server to update information on the edge computer or reverse the edge computer's decision making can result in overall lower system performance. We then perform a series of numerical experiments to understand how changing various parameters affect the problem. We show for the general real-world scenario, the key factor influencing the decision is the ability of the edge computer to detect false alarms. We also show magnitude of loss and ratio of real to false incidents have a linear and logarithmic relationship to the reliability of the edge computer. Collapse

Topics: decision making internet of things human in the loop human computer interaction edge computing

Methods: experiment computational algorithm theoretical contribution simulation

Resource Allocation for Heterogeneous Computing Tasks in Wirelessly Powered MEC-enabled IIOT Systems

2023 | ACM Transactions on Management Information Systems | Citations: 0

Authors: Hu, Yixiang; Deng, Xiaoheng; Zhu, Congxu; Chen, Xuechen; Chi, Laixin

Abstract: Integrating wireless power transfer with mobile edge computing (MEC) has become ... Expand

Abstract: Integrating wireless power transfer with mobile edge computing (MEC) has become a powerful solution for increasingly complicated and dynamic industrial Internet of Things (IIOT) systems. However, the traditional approaches overlooked the heterogeneity of the tasks and the dynamic arrival of energy in wirelessly powered MEC-enabled IIOT systems. In this article, we formulate the problem of maximizing the product of the computing rate and the task execution success rate for heterogeneous tasks. To manage energy harvesting adaptively and select appropriate computing modes, we devise an online resource allocation and computation offloading approach based on deep reinforcement learning. We decompose this approach into two stages: an offloading decision stage and a stopping decision stage. The purpose of the offloading decision stage is to select the computing mode and dynamically allocate the computation round length for each task after learning from the channel state information and the task experience. This stage allows the system to support heterogeneous computing tasks. Subsequently, in the second stage, we adaptively adjust the number of fading slots devoted to energy harvesting in each round in accordance with the status of each fading slot. Simulation results show that our proposed algorithm can better allocate resources for heterogeneous tasks and reduce the ratio of failed tasks and energy consumption when compared with several existing algorithms. Collapse

Topics: resource allocation edge computing decision quality

Methods: computational algorithm reinforcement learning simulation experiment PyTorch

Creating Intelligent Computational Edge through Semantic Mediation

2023 | HICSS | Citations: 0

Authors: Juric, Radmila; Ronchieri, Elisabetta; Suh, Sang

Abstract: This research proposes semantic mediation based on reasoning and the first order ... Expand

Abstract: This research proposes semantic mediation based on reasoning and the first order logic for mediating the best possible configuration of Computational Edge, relevant for software applications which may benefit for running computations with proximity to their data sources. The mediation considers the context in which these applications exist and exploits the semantic of that context for decision making on where computational elements should reside and which data they should use. The application of semantic mediation could address the initiative to accommodate algorithms from predictive and learning technologies, push AI towards computational edges and potentially contribute towards creating a computing continuum. Collapse

Topics: cloud computing mobile system internet of things edge computing artificial intelligence

Methods: ontological modelling computational algorithm literature study literature sample

Introduction to the Minitrack on Intelligent Edge Computing in Pervasive Environments

2023 | HICSS | Citations: 0

Authors: Juric, Radmila; Ronchieri, Elisabetta

Abstract: This is a new minitrack in the software technology track aiming to address the ... Expand

Abstract: This is a new minitrack in the software technology track aiming to address the newest trends in software development, where more power is given to computational edges in accommodating new software technologies and therefore potentially creating an intelligent edge computing paradigm. The submissions to the minitrack were interesting, but we chose only two. They are rather different papers, but they both reflect upon the complexity of modern computing in pervasive spaces, difficulties in maintaining the presence of powerful clouds and attempts to accommodate learning and predictive technologies within modern software solutions across many problem domains.The first paper gives an insight into designing, prototyping and testing pervasive and connected digital twins for edge computing enabled industrial applications. The second comes from collaborative work on creating a flexible computational space between clouds and edges of networks using reasoning and mediation.The paper entitled "Pervasive and Connected Digital Twins for Edge Computing Enabled Industrial Applications" comes from Norway and is the result of collaborative work with academia and industry. Their proposed paradigm allows for the creation of computational models for cloud computing as well as the transmission of data and computational intelligence through analytic platforms. It is very interesting that they exploit the power of digital twins in this case, which has not yet been used in configuring edge computing. This paradigm sits within the architecture, showing the potential convergence of humans, "things" and AI in creating edge intelligence. Their case study shows a Norwegian company, that provides cleaning solutions for entrepreneurs, process industries, mining, and energy operations, being designated as a digital twin. The computational paradigm and suggested architecture in the paper facilitates the proper maintenance and management of infrastructures and facilities of the company. However, the interesting outcome is that there is scope to use this paradigm with "digital twins" in a wide range of applications in future, which opens a new chapter for research and industrial applications of edge computing.The paper entitled "Creating Intelligent Computational Edge through Semantic Mediations" focuses on the constantly changeable computational space between clouds and edges of computer networks, with the goal of creating the best possible configuration of edge computing according to the semantics of the context in which we wish to perform computing. The authors put forward the ideas of mediation on the best possible configuration of computational edges, by using new software technologies which enable reasoning within the given context. This exploration in creating intelligent computational edge is secured by making it independent in terms of deciding through reasoning, which computations we wish to run at the edge and which data the edge needs. These decisions are constantly changeable, and dependent on contexts in which we empower the edge. The authors claim that this might be the first step towards creating a computing continuum which would accommodate complex and data centric applications, currently dominating modern software engineering and shaping the current AI technology. Collapse

Topics: digital twin edge computing

Methods: design artifact

Artificial Intelligence and Blockchain Integration in Business: Trends from a Bibliometric-Content Analysis

2023 | Information Systems Frontiers | Citations: 0

Authors: Kumar, Satish; Lim, Weng Marc; Sivarajah, Uthayasankar; Kaur, Jaspreet

Abstract: Artificial intelligence (AI) and blockchain are the two disruptive technologies ... Expand

Abstract: Artificial intelligence (AI) and blockchain are the two disruptive technologies emerging from the Fourth Industrial Revolution (IR4.0) that have introduced radical shifts in the industry. The amalgamation of AI and blockchain holds tremendous potential to create new business models enabled through digitalization. Although research on the application and convergence of AI and blockchain exists, our understanding of the utility of its integration for business remains fragmented. To address this gap, this study aims to characterize the applications and benefits of integrated AI and blockchain platforms across different verticals of business. Using bibliometric analysis, this study reveals the most influential articles on the subject based on their publications, citations, and importance in the intellectual network. Using content analysis, this study sheds light on the subject’s intellectual structure, which is underpinned by four major thematic clusters focusing on supply chains, healthcare, secure transactions, and finance and accounting. The study concludes with 10 application areas in business that can benefit from these technologies. Collapse

Topics: blockchain artificial intelligence productivity smart contract healthcare data

Methods: bibliometric analysis qualitative content analysis literature sample co-citation analysis machine learning

MaCuDE IS Task Force Phase II Report: Views of Industry Leaders on Big Data Analytics and AI

2023 | Communications of the Association for Information Systems | Citations: 2

Authors: Lyytinen, Kalle; Topi, Heikki; Tang, Jing

Abstract: This paper represents the Phase II report of the Management Curriculum for the ... Expand

Abstract: This paper represents the Phase II report of the Management Curriculum for the Digital Era (MaCuDE) disciplinary task force on information systems (IS). Aligned with the current work of the AIS (Association for Information Systems) and ACM (Association for Computing Machinery), we focus on the current and future industry driven educational needs and requirements posed by big data analytics (BDA), artificial intelligence (AI), machine learning (ML), and related innovations. In this report, we probe and report on the views of industry leaders regarding BDA/AI education needs. We conducted 18 rich semi-structured interviews with a representative sample of industry leaders situated around key changes and issues related to workforce demands in digital transformation and associated educational needs. We performed a grounded theory based analysis of key themes in reported education needs. We note the shifting meaning of AI and BDA phenomena and identify three main organizational level needs for the digital eracapability improvement and transformation, decision-making strategies and tactics, and changes in operations or products -and connect them to three individual professional competencies necessary to deliver them: fundamental environmental competencies, data information and content, and system design competencies. Based on the analysis we outline several novel competency-based IS curriculum recommendations for the master's and undergraduate level IS education.Architecting for cloud: "And so the thing I noticed, a lot of these companies, when their data is growing so big and their database costs are running so high, they think about going towards Open Source Software, Open Source databases to store all the data and, hence, that creates a demand on the workforce to actually know how to use those. "Understanding data structures, architectures, and governance: "So that's his role, and the key requirement for that role again is Data Architecture, data Engineering and Data Management" Building and managing machine learning models: "Second, of course, is strong competency in the use of Machine Learning packages or models. This field is evolving aggressively." "For developing or programming our own AI solutions, our employees have no deep knowledge of mathematical models and how to build mature and value-creating AI applications." IS management and operations: "The talent sort of wars out there are really, really tough, and finding people who understand that, (how to) generate the data, use the data, think about the data, manipulate it are really, really important. I agree. It used to be you had to have people with Project Management. That was critical and that was everyone competed for that, and now it's that plus the data, the Machine Learning."Based on the data collected in this study, we cannot offer a clear answer to the questions of how BDA and AI affect areas that do not belong to the two main areas of the study. Given, however, how quickly BDA and AI have grown and how substantial related resource requirements are in organizations, it is clear that they already have an impact on several competency needs such as programming language selection, selection of BDA and AI tools based on cloud-based solutions, data architecture, data governance, service repertoires available on platform environments, and special characteristic of BDA and AI projects and how they affect education for project management needs and competencies. Summary of Findings Related to Individual CompetenciesGenerally, we observe a strong alignment and continuity between the recent information systems curriculum recommendations and the findings of this study. The data highlights the continued importance of IS core topics covered in all past curricula, such as data management, systems development, IT infrastructure and architecture (including security), and IS (project) management.The results also highlight also multiple areas that need more emphasis and careful design decisions when schools plan and implement their BDA and AI based programs: Integration of business and technology competencies in ways that allow seamless communication with both business and technology experts  A stronger emphasis on multiple life cycle models and concepts related to data management and data science than in the recent model curricula focused on system life cycles  Heightened focus on new dimensions of data analytics competencies (including communication-related elements such a storytelling and visualization)  The growing importance of ethics as a foundation for ensuring privacy and security of BDA and AI models  Need for more skills in statistics  Understanding the novel requirements that AI solutions set up for IT infrastructure  The importance of competencies in AI systems development and deployment. Collapse

Topics: artificial intelligence big data analytical information system data management decision making

Methods: qualitative interview machine learning meta learning field study personal interview

Giving to Get Well: Patients’ Willingness to Manage and Share Health Information on AI-Driven Platforms

2023 | Communications of the Association for Information Systems | Citations: 0

Authors: Raddatz, Nirmalee; Kettinger, William; Coyne, Joshua

Abstract: The digitalization of healthcare makes for the widespread availability of patie ... Expand

Abstract: The digitalization of healthcare makes for the widespread availability of patient-provided data. Artificial Intelligence (AI) relies on this data. In this information-intensive environment, it is imperative to understand the contributing factors of an individual's willingness to manage and share personal health information (PHI). Drawing from the health belief model, we identify the factors that motivate individuals to manage and share their PHI in an AI-driven health platform to obtain its intended benefits. We recognize security risks and present the use of a blockchain database as a representative means of securely managing and controlling an individual's PHI. Data collected from a nationally representative sample of allergy sufferers indicate that the health belief model strongly predicts willingness to share PHI on a personalized AI-supported platform. Our study makes significant contributions by investigating the factors that motivate patients to use an AI-driven health platform to manage their health.Artificial intelligence (AI) advances are taking shape and promising to transform the healthcare industry. According to a study by Accenture, AI in the U.S. healthcare industry is expected to have a compound annual growth rate of 46.1% and reach $95.65 billion by 2028 (Vantage Market Research, 2022). AI can offer a more personalized healthcare experience that optimizes the time it takes to gather information for more efficient decision-making, ultimately allowing better patient care. AI is making its presence known across the medical profession, and medical professionals, researchers, patients, and healthcare providers that do not embrace this technology could be left behind.Ultimately, the success of AI-based platforms for medical purposes depends heavily on individual willingness to use the technology. 1 AI can collect, combine, and analyze large amounts of data from various sources, increasing information-gathering and processing capabilities tremendously (Deane, 2018). This capability offers numerous benefits but also potential risks. For example, according to Rocher, Hendrickx, and de Montjoye (2019), AI can profile, identify, and track individuals by de-anonymizing once anonymized information. Some AI only needs 15 data points about an individual to identify the person. Because of security risks, people may be reluctant to share their personal health information (PHI) with AIbased platforms (Deane, 2018).Achieving the potential benefits of AI from patient platforms will depend on (1) the development of intelligent systems that can collect and analyze vast datasets and (2) the willingness of data owners to share and manage their data with these intelligent systems (Agarwal & Khuntia, 2009). Within healthcare, these conditions are (1) the development of systems that can collect and process PHI and provide or augment diagnostics as/more efficiently and effectively than a traditional doctor-patient profile and (2) patients' ability and willingness to manage and share their own PHI. AI developers work to satisfy the first condition but often take the second condition for granted with the attitude, "If we build it, they will come." However, the first condition is certainly not sufficient to satisfy the second. Without convincing many individuals to share their PHI, an AI-driven health platform cannot serve its purpose. As a result, the question arises as to how to increase individuals' willingness and ability to manage and share their PHI with AI-powered health platforms. Thus, the purpose of our investigation is to discover how to begin to satisfy this second condition by addressing the following questions:Are patients willing to share their PHI with AI-driven health platforms despite security risks? Are patients empowered and ready to manage their PHI, thus, in large part, bringing the control of information to the individual?While numerous studies focus on the design aspects of AI-based predictive models, a significant lack of studies investigate the behavioral aspects pertaining to the usability of these models by patients and clinicians. We adapt Rosenstock's (1974) health belief model (HBM) and expand it to identify the determinants of our test subjects' willingness to share PHI with an AI-driven health platform that uses a blockchain-based database. We apply the HBM to a patient profile in which patients do not solely rely on traditional healthcare provider information channels to get diagnostic and treatment information. Instead, they are empowered to use their PHI to get diagnostic and treatment information from an AI platform. As a result, our application of the HBM moves beyond a patient's willingness to follow a proactive health behavior to considering how patients might proactively use (manage and share) their personal health information to help them receive health-related guidance and insights. This patient information management and information sharing behavior represent a future path of the healthcare industry, and we wish to determine whether this well-regarded health-centered model explains behavior under this new "informationally empowered" patient health profile.Because we care about the practical usability of AI-driven health platforms rather than purely hypothetical usability, we model our research instrument on an actual AI-driven health platform (i.e., doc.ai). 2 This platform uses AI for diagnostic and treatment information and a blockchain-based database to store PHI in 1 We define an AI-based healthcare platform as a software application that (1) collects and stores data for analysis by artificial intelligence (e.g., machine learning) and (2) makes the results of that data analysis available to patients and healthcare professionals in order to guide diagnostic and treatment decisions. We base this definition on an actual AI-based healthcare platform (doc.ai) that we describe in section 2. 2 We provide more specific information regarding this health platform in sections 2 and 3. Collapse

Topics: artificial intelligence self efficacy healthcare data blockchain database system

Methods: survey survey design SEM tool chi squared test machine learning

Theories: health belief model technology acceptance model

Lessons from the Regulation of E-scooters through the MDS Standard: Policy Lessons for Connected Vehicles

2023 | HICSS | Citations: 0

Authors: Rudmark, D.; Sandberg, J.; Watson, R.

Abstract: Connected vehicles generate new data streams that present promising opportuniti ... Expand

Abstract: Connected vehicles generate new data streams that present promising opportunities for policymakers to monitor and learn from events and behavior. To explore what we can learn from how public entities leverage ubiquitous data streams for policy development and enforcement, we draw on a case study of the standard Mobility Data Specification (MDS) and its use by cities to regulate E-scooter operators. Our findings suggest that (1) the richness of real-time data changes the speed of policy revision, (2) data access enables moving some micro-decisions to the edge, and (3) policy will be formulated as fixed or flexible with different amendment rules. Collapse

Topics: video conference edge computing anonymity application programming interface

Methods: qualitative interview theoretical contribution case study single case study personal interview

Theories: institutional theory

PERSONAL DATA PROTECTION RULES! GUIDELINES FOR PRIVACY-FRIENDLY SMART ENERGY SERVICES

2022 | European Conference On Information Systems | Citations: 0

Authors: Gerlach, Jana; Scheunert, Alexandra; Breitner, Michael H.

Abstract: Privacy-friendly processing of personal data is proving to be increasingly chal ... Expand

Abstract: Privacy-friendly processing of personal data is proving to be increasingly challenging in today's energy systems as the amount of data grows. Smart energy services provide value creation and co-creation by processing sensible user data collected from smart meters, smart home devices, storage systems, and renewable energy plants. To address this challenge, we analyze key topics and develop design requirements and design principles for privacy-friendly personal data processing in smart energy services. We identify these key topics through expert interviews, text-mining, and topic modelling techniques based on 149 publications. Following this, we derive our design requirements and principles and evaluate these with experts and an applicability check with three real-world smart energy services. Based on our results and findings, we establish a further research agenda consisting of five specific research directions. Collapse

Topics: privacy legal context General Data Protection Regulation data protection knowledge base

Methods: qualitative interview archival research natural language processing design science topic model