Deploying scalable Vision Transformer applications on mobile and edge devices is constrained by limited memory and computational resources. Existing model development and deployment strategies include distributed computing and inference methods such as federated learning, split computing, collaborative inference and edge-cloud offloading mechanisms. While these strategies have deployment advantages, they fail to optimize memory usage and processing efficiency, resulting in increased energy consumption. This paper optimizes energy consumption by introducing adaptive model partitioning mechanisms and dynamic scaling methods for ViTs such as EfficientViT and TinyViT, adjusting model complexity based on the available computational resources and operating conditions. We implement energy-efficient strategies that minimize inter-layer communication for distributed machine learning across edge devices, thereby reducing energy consumption from data flow and computation. Our evaluations on a series of benchmark models show improvements, including up to a 32.6% reduction in latency and 16.6% energy savings, while maintaining mean average precision sacrifices within 2.5 to 4.5% of baseline models. These results show that our proposal is a practical approach for improving edge AI sustainability and efficiency.

Implementing artificial intelligence (AI) in public settings requires a fundamental transformation of various social and technical aspects within public administration. However, the transformative efforts required for AI integration and use in government remain underexplored. This study introduces the concept of 'AI-augmented government transformation,' building on sociomateriality and sociotechnical theory, and develops a theoretical framework to explore this phenomenon. By applying this framework and drawing insights from expert interviews, we identify the strategic shifts and socio-technical adaptations essential for integrating AI into public administrations. Our analysis highlights the importance of opening the 'black box' of AI to gain a deep understanding of its underlying technologies and their materialities. The findings reveal complex interdependencies between AI materiality and the social and technical systems that public administrations must navigate. Specifically, AI, as a novel materiality, introduces new organizational dynamics, enhances employee capabilities, and alters operational routines and practices. These changes complement technical ones, such as upgrades and advancements in data collection and processing. By investigating the complexities of AI-augmented government transformation, this research offers novel and practical insights for policymakers and practitioners navigating the challenges and opportunities of AI integration.

Organizations increasingly use Generative Artificial Intelligence (AI) to create strategic documents, legislation, and recommendations to support decision-making. Many current AI initiatives are technology-deterministic, whereas technology co-evolves with the social environment, resulting in new applications and situations. This paper presents a novel view of AI governance by organizations from the perspective of complex adaptive systems (CASs). AI is conceptualized as a socio-technological and adaptive system in which people, policies, systems, data, AI, processes, and other elements co-evolve. The CAS lens draws attention to focusing AI governance on the entire organization, taking an outward perspective and considering public values and societal concerns. Although there is no shortage of AI governance instruments, they differ in their effectiveness, and combinations of appropriate mechanisms should be selected to deal with AI's evolving nature and complexity. A major challenge is that no responsibility, and therefore accountability, is taken due to the lack of understanding of the full socio-technological CAS. As such, joint accountability is needed in which involved parties work together.

Vulnerable road users (VRUs), including pedestrians, cyclists, and motorcyclists, account for approximately 50% of road traffic fatalities globally, as per the World Health Organization. In these scenarios, the accuracy and fairness of perception applications used in autonomous driving become critical to reduce such risks. For machine learning models, performing object classification and detection tasks, the focus has been on improving accuracy and enhancing model performance metrics; however, issues such as biases inherited in models, statistical imbalances and disparities within the datasets are often overlooked. Our research addresses these issues by exploring class imbalances among vulnerable road users by focusing on class distribution analysis, evaluating model performance, and bias impact assessment. Using popular CNN models and Vision Transformers (ViTs) with the nuScenes dataset, our performance evaluation shows detection disparities for underrepresented classes. Compared to related work, we focus on metric-specific and cost-sensitive learning for model optimization and bias mitigation, which includes data augmentation and resampling. Using the proposed mitigation approaches, we see improvement in IoU(%) and NDS(%) metrics from 71.3 to 75.6 and 80.6 to 83.7 for the CNN model. Similarly, for ViT, we observe improvement in IoU and NDS metrics from 74.9 to 79.2 and 83.8 to 87.1. This research contributes to developing reliable models while addressing inclusiveness for minority classes in datasets. Code can be accessed at: BiasDet.

Understanding the landscape of privacy protection in governmental systems is crucial for ensuring the trustworthiness of public services and safeguarding citizens' sensitive data from breaches or misuse. Systematic mapping and synthesis of previous research can help identify existing privacy-preserving techniques, assess their effectiveness, and highlight areas for improvement, offering valuable insights for policymakers and practitioners. We aim to conduct a systematic literature review (SLR) of privacy-preserving tools and technologies, focusing on their adoption and governments' challenges. This study also uncovers emerging trends and future research directions, contributing to developing more robust privacy strategies tailored to government needs.Given its extensive reach and government-centric methodology, this evaluation distinguishes itself from previous research. Our work methodically synthesizes privacy-preserving tools and technologies from the distinct perspective of government roles, in contrast to previous assessments that concentrate narrowly on certain technologies or areas. Our findings offer a synthesis of the government's diverse roles in privacy preservation - regulator, enforcer, user, and service provider - and address existing concerns and key privacy-related technologies. Finally, we identify significant research opportunities, such as improving privacy-preserving mechanisms to strengthen the integrity of public services and mitigate the risks of data breaches and misuse.

The metaverse is still in its early stages, and most prototype metaverse platforms fail to fulfill their potential and meet users’ demands. Numerous studies have focused on the success of metaverse applications. However, there has been limited research on the failure of the metaverse services. Based on digital service failure model (DSFM), uses and gratification theory (U&G), and organizational failure diagnosis model (OFDM), an integrated model to investigate the influences of different failure types on citizens’ satisfaction with meta-government services is developed. The model was tested using a survey of 402 responses in China, which were analyzed using a hybrid structural equation modeling artificial neural network (SEM-ANN) approach. The findings indicated that organizational failure exerts the most powerful effect on citizens’ satisfaction. Moreover, information, service, system, and psychological needs failure, negatively influence citizens’ satisfaction. In addition, technology anxiety can weaken the influences of information, system, psychological needs, and organizational failures on citizens’ satisfaction while enhancing the relationship between service failure and citizens’ satisfaction. This research contributes initial and original insights into the failure types of meta-government services and their impacts, addressing a significant gap in the literature.

Vision transformer (ViTs) models have shown higher accuracy, robustness and large volume data processing ability, creating new baselines and references for perception tasks. However, these advantages require large memory and high-performance processors and computing units, which makes model adaptability and deployment challenging within resource-constrained environments such as memory-restricted and battery-powered edge devices. This paper addresses the model deployment challenges by proposing a model approximation approach VI-ViT, for edge deployment using variational inference with mixed precision for processing multi-modalities, such as point clouds and images. Our experimental evaluation on the nuScenes and Waymo datasets show up to 37% and 31% reduction in model parameters and Flops while maintaining a mean average precision of 70.5 compared to 74.8 of the baseline model. This work presents a practical deployment approach for approximating and optimizing Vision Transformers for edge AI applications by balancing model metrics such as parameters, flops, latency, energy consumption, and accuracy, which can easily be adapted to other transformer models and datasets.

Driving assist applications and connected autonomous vehicle systems are supported using AI models and algorithms, which process and analyze heavy data volumes. High-performance computing units and large memory systems support these models, algorithms, and applications, which results in additional onboard energy consumption. The current trend is also towards full electrification of vehicles and increasing connectivity in the vehicular ecosystem to support collaborative and distributed applications using vehicle-edge-cloud computing. However, with the increased focus on model performance and improving the accuracy of these models and applications, the issue of high-performance computing requirements and resulting energy consumption are overlooked. The problem becomes more challenging and complex for resource-constrained edge devices, which are battery-dependent and have limited memory and computing power. This paper proposes components for an adaptive framework to reduce energy consumption by balancing model accuracy. The contributions include proposing and integrating model partition mechanisms, adaptive deployment across edge devices and approximation strategies for the models. By integrating these components, this framework supports energy-aware development across various platforms. The approach offers a sustainable method for computing and communication-oriented applications within the vehicular ecosystem.

In response to unprecedented global urbanization, the smart city concept has emerged, leveraging ICT to enhance municipal efficiency and improve the quality of urban life. The concept of smart energy city (SEC) is closely related to smart cities, however, energy system development in a smart city context is often found eluding certain segments of society, which calls for more attention to inclusion in SEC development. In this paper, the research question is: How can inclusion be effectively integrated into a framework of SEC design? A framework is developed comprising three key principles - energy conservation, energy efficiency, and renewable energy. These principles are aligned with collaboration among stakeholders, smart energy solutions applications, and integration of these solutions. The framework is illustrated using two real-world cases of demonstration projects in the City of Amsterdam, the Netherlands. The paper concludes by presenting several strategies for fostering inclusion in SEC development. They pertain to including utilization of the framework as a guideline to promote inclusion, establishing a clear understanding of inclusion, and involving all relevant stakeholders, including citizens' rights from the project's inception, and fostering transparency regarding the objectives, interests, and individual stakeholders' value.

This is the 11th edition of the workshop series labeled “Web and the City - The Web and Smart Cities”, which started back in Florence in 2015 and kept on taking place every year in conjunction with the WWW conference series. Last year the workshop was held in Singapore. Each year the focus of the workshop is actualized, and this year the workshop focuses on the Responsible web and Responsible AI in cities and communities. In the era of digital twinning, AI, and the metaverse (so-called citiverse for cities), and under the UN 2030 Agenda for sustainable growth and resilience, cities are being transformed into virtual spaces that enable service automation and value generation to their communities and enterprises. Moreover, AI and web intelligence generate new types of opportunities, including generating automated transactions in these virtual spaces, which must preserve human-perceived consequences, fairness, accountability, transparency, and ethics. This workshop aims to focus on how the Web transforms cities into responsibly intelligent virtual environments.

Despite the many promises of open data, numerous challenges inhibit its full potential, such as its poor or inconsistent quality, a lack of complementary assets, and the limited skills of data providers and end-users. Open data intermediaries are instrumental in addressing some of these challenges. They provide specialized resources and capabilities to enhance the supply, flow, or use of open data, or strengthen the relationships among various open data ecosystem (ODE) actors. However, in-depth studies on open data intermediation business models are limited, with most having only offered birds-eye views of those business models. This deficiency has limited our knowledge to develop open data intermediation business models that contribute positively to the ODE. In the geospatial domain, Esri is an important open data intermediary, having been involved in such a role since the 1990s. This article unpacks Esri's open data intermediation business model and analyses its current strengths and weaknesses as well as its potential opportunities and threats to the ODE. Finally, this article recommends factors to consider in developing an open data intermediation business model that supports the sustainability of the ODE.

The digital government research community has developed several publication channels, including journals and conference series to evaluate, refine, and present the outcomes of its work to fellow researchers and the world. Developing and sustaining such channels requires institutional backing with sufficient financial, organizational, and communication capacity. Existing publication channels implement diverse institutional arrangements from commercial publishers and universities to international and non-profit organizations. However, besides institutional support, a key role in operating such channels are journal editors-in-chief and conference program chairs, who link the channels to the research community that supplies them with steady streams of research, and ensure that the channels maintain their relevance to the changing digital government field, ensure the quality of accepted publications, and contribute to the advancement of the field. This paper examines the evolving landscape of publication channels for digital government research, institutional arrangements for developing and sustaining such channels, and the role of the guardians of such channels – editors-in-chief and program chairs. The paper provides the context for a discussion among such guardians on the insights for publishing digital government research.

This is the 10^th edition of the workshop series labeled “AW4City – Web Applications and Smart Cities”, which started back in Florence in 2015 and kept on taking place every year in conjunction with the WWW conference series. Last year the workshop was held in Austin, Texas. The workshop series aims to investigate the Web and Web applications’ role in establishing smart city (SC) promises. The workshop series aim to investigate the role of the Web and of Web applications in SC growth. This year, the workshop focuses on the new era of the Web and web intelligence in cities and communities. In the era of digital twinning and metaverse (so-called citiverse for cities), and under the UN 2030 Agenda for sustainable growth, cities are being transformed into virtual spaces that generate new types of value and new experiences to their citizens and enterprises that can enhance living and offer new opportunities for economic growth. Moreover, AI and web intelligence generate new types of automated transactions in these virtual spaces, while they can utilize data spaces and standardization for optimal data flow. This workshop aims to demonstrate how the Web transforms cities into new virtual environments.

Acceptance by customers is key to the success of shared autonomous vehicles (SAVs). However, only a small group of early technology-savvy customers currently use such vehicles, while the general population does not. Based on the Unified Theory of Acceptance and Use of Technology, Theory of Perceived Risk, and perceived threat of unemployment combined with knowledge of automated vehicles, this research develops an integrated model to investigate the determinants behind the intention to use fully automated taxis. Furthermore, it tested the differences between gender. Through the analysis of 539 samples, the findings showed that performance expectancy, effort expectancy, social influence, and knowledge of automated vehicles positively influence acceptance intention, while perceived safety risk and the perceived threat of unemployment were negatively related to behavioral intention. Moreover, effort expectancy, social influence, and perceived safety risk showed greater influence on females, while knowledge of automated vehicles exerted stronger effects on males.

The digitalization of public administrations can reduce corruption, but many efforts fail. Although the relationship between digitalization and corruption has been investigated, how corruption can be reduced in practice is given hardly any attention. Therefore, we take a different approach by investigating an in-depth case study using documents and interviewing civil servants and citizens. This study found technical and organizational issues blocking the reduction of corruption. Despite encountering technical challenges, most informants agreed that automating administrative processes and making each process step transparent resulted in less corruption. Separating concerns by having different officers conduct each process reduced corruption on the condition that direct communication between immigration officers and passport applicants was blocked. The present study concludes that organizational and technological changes should go hand in hand to reduce corruption. Specifically, the separation of concerns at the organizational and system levels is needed to prevent persons from collaborating and misusing the system. Also, regulatory modifications and suitable governance structures are necessary.

Achieving impact from academic research is a challenging, complex, multifaceted, and interconnected topic with a number of competing priorities and key performance indicators driving the extent and reach of meaningful and measurable benefits from research. Academic researchers are incentivised to publish their research in high-ranking journals and academic conferences but also to demonstrate the impact of their outputs through metrics such as citation counts, altmetrics, policy and practice impacts, and demonstrable institutional decision-making influence. However, academic research has been criticized for: its theoretical emphasis, high degree of complexity, jargon-heavy language, disconnect from industry and societal needs, overly complex and lengthy publishing timeframe, and misalignment between academic and industry objectives. Initiatives such as collaborative research projects and technology transfer offices have attempted to deliver meaningful impact, but significant barriers remain in the identification and evaluation of tangible impact from academic research. This editorial focusses on these aspects to deliver a multi-expert perspective on impact by developing an agenda to deliver more meaningful and demonstrable change to how “impact” can be conceptualized and measured to better align with the aims of academia, industry, and wider society. We present the 4D model - Design, Deliver, Disseminate, and Demonstrate - to provide a structured approach for academia to better align research endeavors with practice and deliver meaningful, tangible benefits to stakeholders.

A metaverse is a three-dimensional virtual space (3D VS) where businesses and individuals worldwide can engage, interact, communicate, transact, and exchange information in real-time through an immersive and collaborative platform. These interactions can create complex relationships influenced by the decision-making processes of businesses. Such complexity can lead to challenges in maintaining relationships, ensuring exclusiveness, preventing misuse, and addressing other ethical issues. Therefore, this study aims to identify ethical principles within the metaverse to guide decision-making and maintain complex relationships between users and businesses. Both qualitative and quantitative data were collected for analysis, and simple random sampling was employed for primary data collection. The empirical analysis was conducted using a mixed-method approach. The study identified four ethical principles that guide complex relationships within the metaverse: business benefit evaluation, fairness, explainability, and reliability principles. These principles positively influence decision-making, which, in turn, positively affects the maintenance of complex relationships within 3D VS.