Going into 2021 many European countries have faced subsequent strong waves of COVID-19 infections that forced governments to reestablish many of the restrictions imposed in the 2020 spring, after a period of relative calm in the summer when the pandemic seemed to be under control. Here, we examine the period prior to the resurgence of the pandemic in autumn, following the interregional mobility due to Italian tourism. Further to a preliminary analysis, we investigate on the spread of the pandemic in Italy during the summer 2020, relating the number of infections with tourism data. In particular, we put the focus on the internal tourism (as international mobility was restricted at the time) between the various Italian regions for the summer holidays. The intuition is that tourist flows, due to the lift of travel restriction in the period when most people went on vacation, has been a determining factor in bringing back the virus, in areas where there were only a handful of cases, thus kick starting the second wave that hit Italy in autumn.
Falls are one of the greatest risks for older adults living at home. They are also one of the biggest factors impacting independence and quality of life. Falling and even the fear of falling can lead to serious physical and mental health issues. Moreover, some afflictions like the self-neglect syndrome make this risk even greater. In this paper, we introduce a system to manage the fall risk before and after it happens, using monocular cameras and an humanoid robot. The proposed system achieves a 75% detection of trip-hazard objects in a heavily cluttered environment, and an 86.11% accuracy in detecting falls after they happened.
Touch screen interactions have been linked to repetitive strain injuries. Their use as an input channel to user interfaces can also have limitations due to screen size or mobility constraints. In this work, we investigate the feasibility of a neck-mounted wearable interface for software interaction, using only subtle head movements. Technologies and configurations for sensoring the neck region, including e-textiles and flex sensors, are considered. The type and placement of the sensors are evaluated. An end-to-end prototype system is developed, which takes the sensor readings from the bespoke neck hardware and wirelessly transmits them to a smart phone, for carrying-out the motion classification and interfacing with applications. The classification accuracy of common head movements using classical machine learning algorithms are evaluated. A classification accuracy of 91% is achieved with data collected from the prototype, for a library of five common head gestures and positions.
In recent years, we have witnessed the proliferation of large amounts of online content generated directly by users with virtually no form of external control, leading to the possible spread of misinformation. The search for effective solutions to this problem is still ongoing, and covers different areas of application, from opinion spam to fake news detection. A more recently investigated scenario, despite the serious risks that incurring disinformation could entail, is that of the online dissemination of health information.
Early approaches in this area focused primarily on user-based studies applied to Web page content. More recently, automated approaches have been developed for both Web pages and social media content, particularly with the advent of the COVID-19 pandemic. These approaches are primarily based on handcrafted features extracted from online content in association with Machine Learning. In this scenario, we focus on Web page content, where there is still room for research to study structural-, content- and context-based features to assess the credibility of Web pages.
Therefore, this work aims to study the effectiveness of such features in association with a deep learning model, starting from an embedded representation of Web pages that has been recently proposed in the context of phishing Web page detection, i.e., Web2Vec.
In the last decade, Online Social Networks (OSNs) have emerged as one of the most disruptive communication platforms with high socio-economic value, used every day by billions of users to connect with friends, colleagues, and any other user with social ties and common interests. Social features are pervasively penetrating in our daily life not only by means of OSNs, but also in other contexts, such as that of vehicular ad-hoc networks. Vehicular Social Networks (VSNs) are enhanced with social features that impact not only on societal services, but especially on connectivity and communication performance. As a result, the coexistence of both OSNs and VSNs can represent a benefit for both two domains, allowing the creation of a unique network framework for connectivity maintenance.
This paper presents a connectivity management technique in a heterogeneous network framework, namely SOLVER, comprised of both OSN and VSN communities for data exchange among nodes belonging to different networks. Connectivity switching is allowed in case of expected connectivity loss, as well as degradation of network performance. Results assess the effectiveness of SOLVER framework as compared to traditional Client-Server and Peer-to-Peer architectural modes.
The decentralization of Social Media applications has gained importance in the last years. In this direction, the HELIOS project has been proposed in order to provide a decentralized platform for social applications. Among these several social applications, the Journalism field is considered an important use case. The main problem which needs to be addressed is preventing the spread of fake news and ensuring the authenticity of the literature for end readers. Furthermore, with the rise of Citizen Journalism, the way of how people can participate to share information is changed. In this paper, we present the HELIOS Citizen Journalism App (CJ) developed in the context of the HELIOS project. The CJ App allows users to contribute content anonymously, based on blockchain technology. After publishing the content via the app, the content is available to publishers for further distribution, on a decentralized P2P and IPFS-based network storage. Furthermore, by making a donation for particular content, the CJ also receives remuneration. In this paper, we show the architecture of the App by describing its components and how it works.
Diseases transmitted by mosquito vectors such as malaria, dengue, and Zika virus are amongst the largest healthcare concerns across the globe today. To tackle such life-threatening diseases, it is vital to evaluate the risk of transmission. Of critical importance in this task is the estimation of vector species populations in an area of interest. Traditional approaches to estimating vector populations involve physically collecting vector samples in traps and manually classifying species, which is highly labor intensive. A promising alternative approach is to classify mosquito species based on the audio signal from their wingbeats. Various traditional machine learning and deep learning models have been developed for such automated acoustic mosquito species classification. But they require data preprocessing and significant computation, limiting their suitability to be deployed on low-cost sensor devices. This paper presents two lightweight deep learning models for mosquito species and sex classification from wingbeat audio signals which are suitable to be deployed on small IoT sensor devices. One model is a 1D CNN and the other combines the 1D CNN with an LSTM model. The models operate directly on a low-sample-rate raw audio signal and thus require no signal preprocessing. Both models achieve a classification accuracy of over 93% on a dataset of recordings of males and females of five species. In addition, we explore the relation between model size and classification accuracy. Through model tuning, we are able to reduce the sizes of both models by approx. 60% while losing only 3% in classification accuracy.
Counting mosquitoes in the wild is a crucial capability for monitoring, prediction, and control of vector-borne diseases. Current approaches are mainly manual, where specially designed mosquito traps or ovitraps are placed in areas of interest and recovered the next day. The counting itself is performed in an entomological laboratory, where individual mosquitoes are classified into species and counted. This process is costly, slow and inefficient. At the same time, mosquito counting is most relevant in tropical and sub-tropical countries, where mosquitoes spread deadly diseases like malaria, yellow fever and dengue fever. Many countries in these regions have relatively weak public health systems and so cannot support large-scale vector counting efforts. In this paper, we present a system architecture and a prototype to count mosquitoes in the wild with an Internet of Things approach. A sensor board is developed to gather audio data, and models are developed to detect, classify, and count mosquito species. Here, we present our prototype and an extensive background study of classifying mosquitoes based on sound recordings and some preliminary results and discussion.
The steady deployment of IoT is paving the road toward concrete implementations of the smart city concept, allowing public/private institutions to sense and model (near) real-time digital replicas of physical processes and environments. This Digital Twin (DT) could be used proactively, as a decision support system, providing insights into possible optimizations of processes in a smart city context.
In this article, we present the design and main building blocks of a DT solution for the Urban Facility Management (UFM) process in the metropolitan area of Bologna, Italy. The Interactive Planning Platform for city District Adaptive Maintenance Operations (IPPODAMO) is a proof of concept solution consisting of a (distributed) multi-layer geographical system, fed with heterogeneous data sources originating from different urban data providers. The data are subject to continuous refinements and algorithmic processes, used to quantify and predict near-to-long term evolution of the urban activity level, exploited for planning purposes, scheduling urban maintenance operations and interventions.
The fifth generation of mobile cellular networks (5G) will bring many improvements to our everyday lives. In particular, the interplay between 5G and IoT technologies will pave the way for the realisation of the Smart City concept, providing humans with better, safer and cleaner cities. In this work, we describe the prototype of a Smart Waste Bin, a connected trash bin that automatically sorts garbage with the help of Convolutional Neural Networks algorithms. The bin intelligence runs on the edge of the 5G network exploiting a Mobile Edge Computing (MEC) architecture, enabling a low-latency and energy efficient operation. We describe the design, the implementation and the experimental validation of the prototype, also discussing possible future application scenarios.
Time series forecasting with additional spatial information has attracted a tremendous amount of attention in recent research, due to its importance in various real-world applications on social studies, such as conflict prediction and pandemic forecasting. Conventional machine learning methods either consider temporal dependencies only, or treat spatial and temporal relations as two separate autoregressive models, namely, space-time autoregressive models. Such methods suffer when it comes to long-term forecasting or predictions for large-scale areas, due to the high nonlinearity and complexity of spatio-temporal data. In this paper, we propose to address these challenges using spatio-temporal graph neural networks. Empirical results on Violence Early Warning System (ViEWS) dataset and U.S. Covid-19 dataset indicate that our method significantly improved performance over the baseline approaches.
Tourism and hospitality have shown to be one of the pioneering sectors in the sharing economy, taking advantages of a collaborative and peer-to-peer market model. In this context, we present ShareCities, a system that allows tourists to exploit a 360° virtual representation of locals' room, customized to include peculiar details related to the host's life/interests/hobbies. Using ShareCities, locals can publish information about themselves and about what to see and do in their locale, while visitors can browse authentic and unmediated information provided by locals, and use the platform to initiate a conversation with them. To evaluate our approach and verify if 360° VR and customization could foster personal connections and affinity, and, eventually, empathy, we carried out an experiment engaging 19 users, obtaining interesting results.
Due to the dynamic nature of changes in various ICT technologies nowadays, the gaps between industry, research, and academia need to be bridged in order to adequately support STEM students towards their future career paths. With the COVID-19 pandemic, and restrictions on access to university premises, an agile transition of both teaching and experimentation was essential, and adjustments in the curriculum were needed more than ever. Therefore, in this paper we present an adaptive and on-demand education framework for engineering students, thereby enabling remote experimentation and adjustments of exercise content to enhance students' learning experience. We present the two types of practical experimentation environments, i.e., cloud and real-life net-working testbed, for performing remote laboratory exercises, as well as the assessment of students' experience that is used as an input for the dynamic adjustments of the exercise content. Our results show that students consider they significantly improved the baseline skills our courses tend to build and strengthen towards preparing students for their future jobs.
A major roadblock in machine learning for healthcare is the inability of data to be shared broadly, due to privacy concerns. Privacy preserving synthetic data generation is increasingly being seen as a solution to this problem. However, since healthcare data often has significant site-specific biases, it has motivated the use of federated learning when the goal is to utilize data from multiple sites for machine learning model training. Here, we introduce FELICIA (FEderated LearnIng with a CentralIzed Adversary), a generative mechanism enabling collaborative learning. It is a generalized extension of the (local) PrivGAN mechanism allowing to take into account the diversity (non-IID) nature of the federated sites. In particular, we show how a site with limited and biased data could benefit from other sites while keeping data from all the sources private. FELICIA works for a large family of Generative Adversarial Networks (GAN) architectures including vanilla and conditional GANs as demonstrated in this work. We show that by using the FELICIA mechanism, a site with a limited amount of images can generate high-quality synthetic images with improved utility, while none of the sites need to provide access to their real data. The sharing happens solely through a central discriminator with access limited to synthetic data. We demonstrate these benefits on several realistic healthcare scenarios using benchmark image datasets (MNIST, CIFAR-10) as well as on medical images for the task of skin lesion classification. We show that the utility of synthetic images generated by FELICIA surpasses that of the data available locally and we demonstrate that it can correct the reduced utility of a biased subgroup within a class.
Lithium-ion battery technologies play a key role in transforming the economy reducing its dependency on fossil fuels. Transportation, manufacturing, and services are being electrified. The European Commission predicts that in Europe everything that can be electrified will be electrified within a decade. The ability to accurate state of charge (SOC) estimation is crucial to ensure the safety of the operation of battery-powered electric devices and to guide users taking behaviors that can extend battery life and re-usability. In this paper, we investigate how machine learning models can predict the SOC of cylindrical Li-Ion batteries considering a variety of cells under different charge-discharge cycles.
This study investigated how staff working at a Danish nursing home experienced, perceived, and used circadian lighting for two years after its installation. The purpose of the installed circadian lighting was to improve the staff and residents' health and comfort. The paper is based on an action research methodology that used interviews, observations, and a questionnaire to investigate 42 staff members' perceived visual comfort, satisfaction with, and perceptions of the usefulness of the circadian lighting. The findings revealed that circadian light was perceived as satisfactory by the staff and was perceived as a more adequate light for work than the existing lighting system. Being able to adjust the lighting was perceived as important by staff for maintaining visibility, setting the lighting depending on the activities, and meeting residents' needs. This paper demonstrates the value of applying mixed methods when analyzing subjective assessment of light and visual comfort. We present an alternative card sorting method for studying perceptions of a 24-hour lighting application. Finally, the study demonstrates the value of evaluating the lighting with end-users after two years in use to improve future lighting installations and to adjust the current installation.
Despite the increasing concerns on sustainability issues worldwide, IoT devices are still being designed unsustainably, and often end up as e-waste in landfill after a very short lifespan. This state of affairs is alarming, as we expect hundreds of billion connected devices in a few years, and calls for solutions to help maximizing the usable lifetime of IoT systems. In this paper, we review this problem in detail, and argue that legacy devices using outdated wireless technologies could make use of cross-technology communication to directly interact with newer IoT products, thereby increasing their durability. We also summarize our recent efforts in this domain and highlight how they could help in designing sustainable IoT systems.
Every year, in Europe alone, hundreds of workers die by falling from high height. This number could be greatly reduced by means of better training and quick detection of individuals with issues toward work at height. Workers proving to be less suited for the job can be subject to more intensive training or recruited for different positions. Unfortunately, the early detection of workers unsuited for working at height involves specialized personnel and expensive equipment to recreate a stressful environment. In this paper we propose a methodology to predict fear of heights by means of a virtual reality environment. We demonstrate that a 3D virtual environment is feasible for the prediction and give guidelines about meaningful physiological parameters useful for detection.
Cloud computing has become one of the most important technologies that have changed the traditional application development and operation (DevOps) lifecycle. However, current cloud software DevOps often faces the following key challenges: 1) selecting the best fitting service providers, customizing services and planning capacities for large-scale distributed applications; 2) guaranteeing high-quality and trustworthy service level agreements (SLAs) among multiple service providers; 3) enhancing the interoperability of cloud services across providers; 4) designing incentive model effectively among players. In this study, a framework called AWESOME is proposed to build a decentralized cloud marketplace and to address the above challenges. The proposed framework contains four subsystems including a customizable auction model, an incentive witness mechanism, and a social behavior-based simulator as one automated framework. We also provide a proof of concept to demonstrate that the AWESOME framework is feasible.
The ADAPT project started during the most critical phase of the COVID-19 outbreak in Europe when the demand for Personal Protective Equipment (PPE) from each country's healthcare system surpassed national stock amounts. Due to national shutdowns, reduced transport logistics, and containment measures on the federal and provincial levels, the authorities could not meet the rising demand from the health care system on the PPE equipment. Fortunately, the PPE production capacities in China have regained (and expanded) their available capacities through which Austria now can get the demand of PPE to protect its citizens.
ADAPT develops an adaptive and autonomous decision-making network to support the involved stakeholders along the PPE supply chain to save and protect human lives. The ADAPT decentralized blockchain platform optimizes supply, demand, and transport capacities between China and Austria with transparent, real-time certification checks on equipment, production documentation, and intelligent decision-making capabilities at all levels of this multidimensional logistic problem.
Recently, a new generation of P2P systems capable of addressing data integrity and authenticity has emerged for the development of new applications for a "more" decentralized Internet, i.e., Distributed Ledger Technologies (DLT) and Decentralized File Systems (DFS). However, these technologies still have some unanswered issues, mostly related to data lookup and discovery. In this paper, first, we propose a Distributed Hash Table (DHT) system that efficiently manages decentralized keyword-based queries executed on data stored in DFS. Through a hypercube logical layout, queries are efficiently routed among the network, where each node is responsible for a specific keywords set and the related contents. Second, we provide a framework for the governance of the above network, based on a Decentralized Autonomous Organization (DAO) implementation. We show how the use of smart contracts enables organizational decision making and rewards for nodes that have actively contributed to the DHT. Finally, we provide experimental validation of an implementation of our proposal, where the execution of the same protocol for different logical nodes of the hypercube allows us to evaluate the efficiency of communication within the network.
The rising of online social platforms makes large volumes of data about social relationships and interactions available to the research community. In the varied ecosystem of techno-social platforms, blockchain-based online social networks - BOSNs - are gaining momentum since the underlying blockchain offers data validation, data storage, and data decentralization. As data sources, BOSNs provide high-resolution temporal data about the evolution of the social network and on the interactions of users with the platform services. In this study, we focus on a few temporal characteristics, by analyzing the dynamics of the link creation process and the claiming of rewards in the BOSN Steemit. We model blockchain data as a temporal directed network from which we extract the time series characterizing link creation and reward claims. Adopting a user-centric approach, we evaluate the heterogeneity of the time series through the inter-event time distribution, the burstiness, the bursty train size distribution, and the fitting of inter-event times by power law models. The outcomes of the analysis highlight that the above processes show bursty traits typical of human dynamics. However, the two aspects present a few differences concerning the types of models describing their behavior and the time scale of their bursty nature. To sum up, the creation of new relationships and the reward claim dynamics ask for specific models able to reproduce their general bursty traits but taking into account their specificities and relations with other services and mechanisms offered by BOSN platforms.
The worldwide spread of the COVID-19 disease has required the adoption of restrictions to protect public health. These rules have heavily changed the access to working places, damaging in-person business, which is still the core of global and local economies. From large corporations to small companies the main challenge is to keep working places open and functional to maintain productivity and efficiency standard under the application of the anti-Covid-19 measures. This paper describes the design, prototyping work, simulations and test procedures of a low-cost system designed to monitor the fluxes of people and air quality of target areas (e.g., company's offices) to warn about the exposition to risky condition and to model crowd flows around a building to advise virtuous behaviour and preserving safe working conditions. To do that, the system merges the data produced from different suits of sensors to create safe working environments and keeping in person business running.
The COVID-19 pandemic has been one of the biggest health crises in recent memory. According to leading scientists, face masks and maintaining six feet of social distancing are the most substantial protections to limit the virus's spread. Experimental data on face mask usage in the US is limited and has not been studied in scale. Thus, an understanding of population compliance with mask recommendations may be helpful in current and future pandemic situations. Knowledge of mask usage will help researchers answer many questions about the spread in various regions. One way to understand mask usage is by monitoring human behavior through publicly available webcams. Recently, researchers have come up with abundant research on face mask detection and recognition but their experiments are performed on datasets that do not reflect real-world complexity. In this paper, we propose a new webcam-based real-world face-mask detection dataset of over 1TB of images collected across different regions of the United States, and we implement state-of-the-art object detection algorithms to understand their effectiveness in such a real-world application.
We describe the work behind a privacy-preserving, crowdsensing approach that promotes social distancing upon the return of students to University. Our main motivation is enabling visualizations that predict room occupancy based on the number of connected devices to particular access points, via anonymous reports about these predictions, and via an unenforced booking system that allows users to communicate their intents about room use.
In crises, citizens show changes in their information behavior, which is mediated by trust in sources, personal relations, online and offline news outlets and information and communication technologies such as apps and social media. Through a repeated one-week survey with closed and open questions of German citizens during the beginning of the COVID-19 pandemic, this study examines citizens' perceptions of information responsibilities, their satisfaction with the fulfillment of these responsibilities and their wishes for improving the information flow. The study shows that the dynamism of the crisis and the federally varying strategies burden citizens who perceive an obligation to stay informed, but view agencies as responsible for making information readily available. The study contributes a deeper understanding of citizens' needs in crises and discusses implications for design of communication tools for dynamic situations that reduce information overload while fulfilling citizens' desire to stay informed.
The aim of this study was to provide engagement with and awareness of the plastic crisis through a serious game, as part of the UN Sustainable Development Goal 12 by targeting students ages 17-21. The study is based on 32 participants, who were included in a formative evaluation within two evaluations (E1 and E2). E1 included an evaluation of 10 university students. E2 consisted of 22 students from a Danish gymnasium, where the game was part a project related to the plastic crisis. Both evaluations followed the same self-reporting methodology, which used a questionnaire and interviews. The findings reveal that most of the participants were engaged and immersed in the serious game, and that they had a high attentional focus. The findings also revealed that the game needs to be improved in terms of usability, aesthetic appeal, and the implementation of additional learning objectives. This study's novelty derives from its measurement of high engagement and immersion in a rather simple platform game about the plastic crisis. Furthermore, this study is intended to emphasize important elements when designing serious games, including content, learning objectives, and the quality of the game's title.
With the increase in the number of people aged 60 years and older, it becomes essential to promote active ageing. Adequate cognitive training activity can preserve older adults' health conditions, allowing them to live autonomously longer. However, older people often have difficulties with the use of digital technologies and risk being excluded from these. For this reason, this work proposes a solution that introduces older adults to new technologies while maintaining a familiar appearance to them. The solution proposed is Micogito, a web serious game that exploits the book metaphor for cognitive stimulation of older adults. It supports tasks which replicate daily living activities, and has been conceived, and designed to stimulate the most implicated cognitive functions in ageing and particularly in mild cognitive impairment. In its design and implementation, guidelines useful to make web applications accessible and usable for the target users were considered. We also report on a first user test with older adults, to assess the effectiveness of this approach, which provided useful and encouraging feedback, and discuss some lessons learnt.
Several neurodevelopmental disorders, such as autism spectrum disorder (ASD), attention-deficit/hyperactivity disorder (ADHD), and intellectual developmental disorders (including Down syndrome), are characterized by deficits in motor coordination, including visual-motor coordination. Deficits in visual-motor coordination can hinder an individual's ability to perform academic activities and activities of daily living hindering their independence. Exergames are a potential tool for supporting the cognitive and motor skills of children with neurodevelopmental disorders. This work presents the design and implementation of Move&Learn, and adaptive exergame to support cognitive and motor coordination skills of children with neurodevelopmental disorders. Move&Learn helps children with neurodevelopmental disorders by executing visual-motor coordination exercises using the players' upper limbs. The exergame can adapt the gameplay according to the players' performance. We validated Move&Learn through a focus group with nine participants, including psychotherapists, parents, a neuropsychologist, and an applied behavior analyst. The results indicate that Move&Learn was perceived as useful and easy to use by the participants. A set of improvements are discussed as future work resulting from the design validation.
Digital games can be designed and used to improve many specific skills, related to both cognitive and sensory-motor goals. A set of novel mini-games explicitly aimed to children with severe visual impairments are presented. The games are played by moving within the range of a large-scale interactive environment, i.e. a floor portion placed under a motion capture system which allows the tracking of one or more people. It is then possible to link the players movements to audio and graphic output, producing meaningful interactions. The games are the results of a design process that involved computer engineers and a multidisciplinary therapy team from the R. Hollman Foundation (Padova, Italy). A set of guidelines useful to designers and developers are presented as a result of the co-design process.
This study investigates the impact of combining virtual environment with a social VRobot as a novel approach to train the social skills of children with high-functioning autism (HFA). A 3D social virtual robot is used in a non-immersive (desktop) virtual environment to enhance the social skills of children with HFA through a social skills training program guided by a parent or a teacher. The motivation of this research is to provide a tool that can be widely accessible, cheap, and easily used by parents and teachers either at home or school. The training program targets three social skills: imitation, emotion recognition, and expression, and intransitive gestures skills. Experimental sessions were conducted with 15 children with HFA (4-12 years) both online and on-site. The participants were taught to recognize six basic emotions and 11 intransitive gestures (Phase I), to imitate these emotions and gestures (Phase II), and to produce them inappropriate social contexts (Phase III). Across all the three phases for each skill, significant differences were found between the pre-test, post-test, and follow-up test results.
Real-time physiological data collection and analysis play a central role in modern well-being applications. Personalized classifiers and detectors have been shown to outperform general classifiers in many contexts. However, building effective personalized classifiers in everyday settings - as opposed to controlled settings - necessitates the online collection of a labeled dataset by interacting with the user. This need leads to several challenges, ranging from building an effective system for the collection of the signals and labels, to developing strategies to interact with the user and building a dataset that represents the many user contexts that occur in daily life. Based on a stress detection use case, this paper (1) builds a system for the real-time collection and analysis of photoplethysmogram, acceleration, gyroscope, and gravity data from a wearable sensor, as well as self-reported stress labels based on Ecological Momentary Assessment (EMA), and (2) collects and analyzes a dataset to extract statistics of users' response to queries and the quality of the collected signals as a function of the context, here defined as the user's activity and the time of the day.
Access to accurate information is paramount to sound decision-making, improves societal trust in governmental and civil society structures and builds civic unity. Disinformation impacts societies' well-being, physical and mental health, financial welfare and even the ability to withstand external aggression. Changes in the way that individuals acquire information resulted in new approaches to influence society. In politics, disinformation is a growing concern, threatening the credibility of democratic processes.
This paper investigated the degree to which misinformation was present during recent US election years by analyzing data on two major social media platforms. The findings show that such events are connected to increases in disinformation activity despite efforts taken by social media companies.
Recent progress in machine learning frameworks makes it now possible to run an inference with sophisticated machine learning models on tiny microcontrollers. Model training, however, is typically done separately on powerful computers. There, the training process has abundant CPU and memory resources to process the stored datasets. In this work, we explore a different approach: training the model directly on the microcontroller. We implement this approach for a keyword spotting task. Then, we extend the training process using federated learning among microcontrollers. Our experiments with model training show an overall trend of decreasing loss with the increase of training epochs.
Coastal lagoons are ecosystems with significant environmental and socio-economic value. However, these natural systems are especially vulnerable to climatic and anthropogenic pressures, such as intensive agriculture and extensive urbanization as a consequence of the tourist development. Despite the vulnerability and complexity of these ecosystems, there has been limited development of novel techniques which can provide real-time monitoring, analysis and management of these critical resources. Beyond being useful for policy-making procedures at multiple levels of granularity, such tools can increase local and citizen awareness of environmental impacts. This paper introduces the key concept behind the SMARTLAGOON project, which intends to develop a digital twin to build a systemic understanding of the socio-environmental inter-relationships affecting coastal lagoons and their ecosystem. Particularly, we focus on the main research activities carried out since the beginning of the project (January 1st, 2021), which are mainly based on initial mapping and reporting of the main stakeholders' needs and wishes in relation to the technological products to be developed in SMARTLAGOON. This will onset the initial development pathway, while additional stakeholder inputs during the project lifetime will help to further explore and prioritize the key developments, thus maximizing the relevance and impact of the Mar Menor's digital twin.
New challenges such as climate change and sustainability arise in society influencing not only environmental issues but human's health directly. To face these new challenges IT technologies and their application to environmental intelligent monitoring become into a powerful tool to set new policies and blueprints to contribute to social good. In the new H2020 project, WatchPlant will provide new tools for environmental intelligence monitoring by the use of plants as "well-being" sensors of the environment they inhabit. This will be possible by equipping plants with a net of communicated wireless self-powered sensors, coupled with artificial intelligence (AI) to become plants into "biohybrid organisms" to test exposure-effects links between plant and the environment. It will become plants into a new tool to be aware of the environment status in a very early stage towards in-situ monitoring. Additionally, the system is devoted to be sustainable and energy-efficient thanks to the use of clean energy sources such as solar cells and a enzymatic biofuel cell (BFC) together with its self-deployment, self-awareness, adaptation, artificial evolution and the AI capabilities. In this concept paper, WatchPlant will envision how to face this challenge by joining interdisciplinary efforts to access the plant sap for energy harvesting and sensing purposes and become plants into "biohybrid organisms" to benefit social good in terms of environmental monitoring in urban scenarios.
Natural radioactivity in the marine environment has been present since the Earth's formation, while artificial radionuclides were introduced into the oceans in 1944. More recent direct sources exist that feed the oceans, such as low-level liquid discharges from reprocessing plants, large-scale releases due to disasters (e.g. Fukushima hit by the tsunami in 2011), and smaller-scale radiological events. Exploration of submarine environments should consider the existence of radioactivity in terms of its short- and long-term impact on marine and coastal ecosystems, also in correlation to natural hazards, such as seismic activity over submarine faults or activity of hydrothermal vent fields near the seabed. Significantly undersampled in oceans, radioactivity poses real risks to marine ecosystems and human population, urging for detailed, data-driven modeling.
RAMONES is a new H2020-EU FET Proactive Project  aiming to offer new and efficient solutions for in in situ, continuous, long-term monitoring of radioactivity in harsh subsea environments. A new generation of submarine radiation sensing instruments, assisted by state-of-the-art (SoA) robotics and artificial intelligence (AI) will be developed towards understanding radiation related risks near and far from coastal areas, while providing data towards shaping new policies and guidelines for environmental sustainability, economic growth and human health, offering a framework for defining future environmental intelligence guidelines and practices.
The main ambition is to lay a radical new path to close the existing marine radioactivity under-sampling gap and foster new interdisciplinary research in threatened natural deep-sea ecosystems. RAMONES will invest a significant effort to provide tools for long-term, rapid deployments, propose new robotics and AI-driven supported methodologies, and offer scaled-up solutions to researchers, policy makers and communities. RAMONES will combine SoA equipment from various disciplines and advanced modeling in fine synergy, and design new and effective approaches for the marine environment to provide efficient response to natural and man-made hazards, shaping future policies for the global population.
Understanding and monitoring natural ecosystems is necessary for an efficient implementation of sustainable strategies to tackle climate and environmental-related challenges, such as: protect and improve the quality of air, water, and soil; safeguard species biodiversity; and effectively manage natural resources. A longstanding challenge for environmental monitoring is the low spatial and temporal resolution of available data for many regions. Also, new approaches for the design of sustainable technologies is urgently needed to reduce current problems related to energy costs and e-waste produced.
With this in mind, the EU-funded FET Proactive Environmental Intelligence project "I-Seed" (Grant Agreement n. 101017940, https://www.iseedproject. eu/) targets towards the development of a radically simplified and environmentally friendly approach for analysing and monitoring topsoil and air. Specifically, I-Seed aims at developing a new generation of self-deployable and biodegradable soft miniaturized robots, inspired by the morphology and dispersion abilities of plant seeds, able to perform a low-cost, environmentally responsible, and in-situ detection. The natural functional mechanisms of seeds dispersal offer a rich source of robust, highly adaptive, mass and energy efficient mechanisms, and behavioural and morphological intelligence, which can be selected and implemented for advanced, but simple, technological inventions. I-Seed robots are conceived as unique in their movement abilities because inspired by passive mechanisms and materials of natural seeds, and unique in their environmentally friendly design because made of all biodegradable components. Sensing is based on a chemical transduction mechanism in a stimulus-responsive sensor material with fluorescence-based optical readout, which can be read via one or more drones equipped with fluorescent LiDAR technology and a software able to perform a real time georeferencing of data.
The I-Seed robotic ecosystem is envisioned to be used for collecting environmental data in-situ with high spatial and temporal resolution across large remote areas where no monitoring data are available, and thus for extending current environmental sensor frameworks and data analysis systems.
Intelligence is the ability to learn, understand and thus manage new or trying situations through reasoning (inferences based on facts or premises). Environmental Intelligence brings together multiple data streams (facts) from ground-based, satellite and citizen sources with cutting-edge hardware, software and analytical technology employing human reasoning and machine learning to better understand and manage the environment.
The EC H2020 ReSET project (Restarting Economy in Support of Environment through Technology) funded by the European Union's Horizon 2020 FET Proactive Programme under grant agreement No 101017857, brings together environmental scientists, social scientists, informatics specialists and stakeholders from five European countries to develop state of the art investment policy support systems. These combine the best available earth observation, crowdsourced and field-monitored data with sophisticated spatial policy support systems for biophysical and social processes. Harnessing combined machine and human intelligence, we seek to to understand best-bet options for 'build back better' investments that maximise environmental, economic and employment benefits.
We are working at a series of demonstration sites in Europe where 'build back better' investments are active: Thames Gateway, OxCam Arc and Strand Aldwych in UK; Carasuhat Wetlands in Romania; Castilla Leon and Rivas VaciaMadrid in Spain and Bologna in Italy.
Proposed investments include urban greening and traffic management to reduce air pollution and thermal extremes (Strand Aldwych, Bologna, Rivas Vaciamadrid); Natural Flood Management (Thames Gateway, OxCam Arc, Castilla-Leon), land use zoning for low impact tourism (Carasuhat) and green-grey approaches to flood and drought management (Castilla Leon).
We bring together new hardware technologies enabling low-cost, distributed, IoT environmental monitoring using the FreeStation.org platform with further developments of our widely used policy support systems CotingNature and Eco:Actuary and enhanced activity and agent-based modelling in the Metronamica modelling framework. This is to better understand current environmental conditions in the areas proposed for investment and to simulate the impact of investment alternatives (business as usual grey, blended grey-green and fully green) on environment, economy and employment in the ReSET investment policy support system.
Through this work, we tackle some key challenges of operationalizing environmental intelligence discussed here:
• technology as an enabler of research and innovation rather than the key focus of research
• live integration of complex data streams
• ensuring usability and ease of use through co-design
• scalability and relevance to a range of investment types and settings
• reducing costs, enabling local maintenance and ensuring accessibility and legacy
In recent years, the focus on sustainability has grown by everyone, including policymakers, companies, and consumers. In this perspective, recycling plays an important role because it allows to reduce the amount of waste to be disposed of, at the same time reducing the need for raw materials. This paper presents ScanBage, a web application designed and developed to support users in separating waste collection. It exploits two machine learning algorithms to automatically classify garbage categories and it employs Gamification elements with the aim of increasing user involvement.
Children are increasingly using the internet nowadays. While internet use exposes children to various privacy and security risks, few studies have examined how parents perceive and address their children's cybersecurity risks. To address this gap, we conducted a qualitative study with 25 parents living in Norway with children aged between 10 to 15. We conducted semi-structured interviews with the parents and performed a thematic analysis of the interview data. The results of this paper include a list of cybersecurity awareness needs for children from a parental perspective, a list of learning resources for children, and a list of challenges for parents to ensure cybersecurity at home. Our results are useful for developers and educators in developing cybersecurity solutions for children. Future research should focus on defining cybersecurity theories and practices that contribute to children's and parents' awareness about cybersecurity risks, needs, and solutions.
We describe how we redesigned, because of the 2020 COVID-19 pandemic, the CS1 course for Math undergraduates to be held online yet reflecting the face-to-face (F2F) experience as much as possible. We present the course structure, the IT tools we used, and the strategies we implemented to preserve the benefits of a synchronous experience. We discuss the positive and negative aspects that emerged from the students' opinion qualitative analysis. We use the COI framework as a lens to explain what worked, what did not, and what can be improved to strengthen the perception of a F2F experience and mitigate the "presence paradox" we found: despite students being enthusiastic about the online format, most would still prefer a F2F course.
Smart homes research is now entering a phase of real deployment and of early commercialization. For the type of smart homes used to monitor the daily life of residents, activity recognition is one of the key artificial intelligence components necessary. In labs, it is mostly based on machine learning methods, but in real deployments, due to the difficulty to build labeled datasets, it still usually depends largely on logical systems and inference rules. In this work, we try to leverage generalizable fuzzy features to evaluate the quality of the label inferred by commonsense inference. The fuzzy rules are built from annotated instances in CASAS's dataset and by transferring them to our own infrastructure. The data exploited include 11 of our deployed smart homes and shows promising results. Our experiments shows that it is likely possible to exploit those rules to evaluate the quality of our data labeling.
Smart traffic management is essential for smart cities. Detection and classification of traffic signs is essential in autonomous driving and in helping with traffic navigation in general. Automated recognition of street parking signs, however, is a more challenging task, due to their small size and huge diversity. Identification of street parking signs is not only helpful in autonomous driving but also in notifying drivers of available parking spaces ahead, significantly alleviating traffic congestion. In this paper, we introduce a new street parking sign detection and classification method that is based on a deep learning network that receives a video feed from a car camera to accurately detect and classify different street parking signs in real-time. Performance evaluations showed that our model achieves mean average precision of 98.56%.
Agriculture including aquaculture has been changing through multiple technological transformations in recent years. The Internet of Things (IoT) and Artificial Intelligence (AI) are providing remarkable technological innovations on fish farming. In this research, we present an automated IoT and AI-based system to improve fish farming. The proposed system uses multiple sensors to measure in real-time water quality chemical parameters such as: temperature, pH, turbidity, electrical conductivity, total dissolved solids, etc., from the fish pond and send them on a cloud database to allow fish farmers to access them in realtime with their devices (mobile phone, PC, tablets). The system contains three web applications which fish farmers can use. The first web application enables farmers with realtime visualizations of sensors data, issues alerts and remote pumps controls. Fish farmers can use the second web application for fish disease detection and to receive suggestions for diseases' care. This would help to classify two fish diseases which are: Epizootic Ulcerative Syndrome(EUS), and Ichthyophthirus(Ich). The third web application is a digital community platform for knowledge sharing, capacity building, market opportunities and collaboration among fish farmers. Our system can help reduce human efforts, reinforce capacity building, increase fish production and market opportunities for fish farmers.
Covid-19 has brought with it an onslaught of information for the public, some true and some false, across virtually every platform. For an individual, the task of sifting through the deluge for reliable, accurate facts is significant and potentially off-putting. This matters since fundamentally, containment of the pandemic relies on individuals' compliance with public health measures and their understanding of the need for them, and any barrier to this, including misinformation, can have profoundly negative effects. In this paper we present a conversational AI system which tackles misinformation using a two-pronged approach: firstly, by giving users easy, Natural Language access via speech or text to concise, reliable information synthesised from multiple authoritative sources; and secondly, by directly rebutting commonly circulated myths surrounding coronavirus. The initial system is targeted at staff and students of a University, but has the potential for wide applicability. In tests of the system's Natural Language Understanding (NLU) we achieve an F1-score of 0.906. We also discuss current research challenges in the area of conversational Natural Language interfaces for health information.
Source trustworthiness can help discerning reliable and truthful information. We offer a computable model for the dynamic assessment of sources trustworthiness based on their popularity, knowledge-ability, and reputation. We apply it to the debate among medical experts in Italy during three distinct phases of the SARS-CoV-19 pandemic, and validate it against a dataset of newspaper articles. The model shows promising results in the analysis of expert debates their impact on public opinion.
In the past couple of decades, the topic of ICT for development (ICTD) has received significant attention and across the globe, researchers, educators, and developers have engaged in projects to study, design, and apply technology to improve education and learning. The expectation that use of information and communication technologies (ICT) will improve education and learning, a critical social good, throughout the world have never been higher. Currently, there is no systematic review of existing literature to better understand what areas have been addressed, their impact, and what are the lessons learned for future work. This paper presents review of work on education and learning with ICT. We focus on the ICTD literature as it has a mature body of work and engages directly with the goal of doing good with technology. Through this review I illustrate the diversity of the work that has been undertaken on the topic and outline five areas of to guide future contributions: 1) appreciating the diversity of the domain; 2) taking a long-tern view of technology intervention; 3) understanding contexts and perspectives; 4) preparing for unintended consequences; and, 5) aiming for power parity between dominant stakeholders and intended beneficiaries. I also show that not all efforts have had a positive outcomes and many design and intervention efforts have been critiqued.
The rapid technological evolution of the last years motivated students to develop competencies and capabilities that will prepare them for an unknown future of the 21st century. In this context, teachers intend to optimise the process of learning and make it more dynamic and exciting by introducing gamification. Thus, this paper focuses on a data-driven assessment of geometry competencies, which are essential for developing problem-solving and higher-order thinking skills. We explored them in the domain of knowledge inference, whose primary goal is to predict or measure the students' knowledge over questions as they interact with a learning platform at a specific time. Hence, the main goal of the current paper is to compare several well-known algorithms applied to the data of a geometry game named Shadowspect in order to predict students' performance in terms of classifier metrics such as Area Under Curve (AUC), accuracy, and F1 score. We found Elo to be the algorithm with the best prediction power. However, the rest of the algorithms also showed decent results, and, therefore, we can conclude that all the algorithms hold the potential to measure and estimate the actual knowledge of students. In turn, this means that they can be applied in formal education to improve teaching, learning, organisational efficiency and, as a consequence, this can serve as a basement for a change in the system.
Building bridges to facilitate communication between deaf and hearing people is a pending issue that can no longer be postponed in a technologically advanced society. In this work, we propose a multiplatform service that allows gradually building a collaborative platform to bring the interests of deaf and hearing people closer together. The proposal consists of a client-server architecture with two main parts: a multiplatform interface for video capture and playback, and a sign language recognizer system composed of a skeleton data extractor and a Graph Convolutional Neural Network. The result is a proof-of-concept of the first thematic dictionary accessible through sign language.
In the wake of a disaster, when all communication infrastructure has been damaged, it is critical to find a way to maintain communications to disseminate critical information collected by the rescuers. Even though many proposals and systems have surfaced, the critical drawback found in these systems is the communication range. Here we propose a solution for communication in post-disaster rescue operations based on LoRa, a long-range, low power technology. Using the concepts of Opportunistic Networks, we employ the epidemic forwarding protocol to disseminate data between the rescuers as well as other external parties. In this work, we present our solution, the implementation in a commercially available LoRa platform, built an experimental setup, identified a set of test cases, and evaluated the performance based on these test cases. Based on the initial results, we show that our proposal is a viable solution to be used by rescuers. We intend to work further on this solution to improve it by identifying optimum configurations for its best performance in realistic post-disaster rescue operations.
Child trafficking is a serious problem around the world. Every year there are more than 4 million victims of child trafficking around the world, many of them for the purposes of child sexual exploitation. In collaboration with UK Police and a non-profit focused on child abuse prevention, Global Emancipation Network, we developed a proof-of-concept machine learning pipeline to aid the identification of children from intercepted images. In this work, we focus on images that contain children wearing school uniforms to identify the school of origin. In the absence of a machine learning pipeline, this hugely time consuming and labor intensive task is manually conducted by law enforcement personnel. Thus, by automating aspects of the school identification process, we hope to significantly impact the speed of this portion of child identification. Our proposed pipeline consists of two machine learning models: i) to identify whether an image of a child contains a school uniform in it, and ii) identification of attributes of different school uniform items (such as color/texture of shirts, sweaters, blazers etc.). We describe the data collection, labeling, model development and validation process, along with strategies for efficient searching of schools using the model predictions.
Managing and sharing cultural heritages also in supranational and multi-literate contexts is a very hot research topic. In this paper we discuss the research we are conducting in the DigitalMaktaba project, presenting the first steps for designing an innovative workflow and tool for the automatic extraction of knowledge from documents written in multiple non-Latin languages (Arabic, Persian and Azerbaijani languages). The tool leverages different OCR, text processing techniques and linguistic corpora in order to provide both a highly accurate extracted text and a rich metadata content, overcoming typical limitations of current state-of-the-art systems; this will enable in the near future the development of an automatic cataloguer which we hope will ultimately help in better preserving and conserving culture in such a demanding scenario.
The advancements in terms of networking, image resolution, computer vision (CV) and mobile cloud computing performances are transforming Mobile Augmented Reality (MAR) into a technology which may be put to good use in a variety of everyday contexts. To make this point, we here consider an artisanal craft rooted back into the past, key locksmithing, and show how today MAR capabilities may simplify and ameliorate the performances of such ancient trade. To this aim, we introduce the requirements posed by such craft and present a MAR-based workflow which may be implemented to support and speed its execution. This could also impact the everyday lives of ordinary citizens, since it pose the bases of remote locksmithing activities.
The tremendous CoVid-19 outbreak has had a significant impact on the lives of people anywhere in the world. Several approaches have been presented with the aim of helping in limiting and mitigating the effect of CoVid-19 infections. This paper proposes a system that exploits Internet of Things (IoT) and Data Visualization to monitor the flow of people in buildings, with the aim of providing policymakers with a tool that visualize and highlight critical issues. The case study considered is a Smart University Campus located in Cesena, belonging to the University of Bologna.
Technologies have changed many different aspects of people's life and the recent CoVid-19 pandemic proved that education is not an exception. But technologies in education go beyond the simple use of video lectures: technologies might be exploited to improve personal learning. In this paper, we focus on the dropout of studies, a global phenomenon that artificial intelligence techniques are trying to ameliorate. Here, we investigate whether data related to the consumption of video lectures might improve the students' dropout prediction. We consider first-year students enrolled in our Department and we characterize them with personal, scholastic, academic and technological features. Then, we measure the performance of three machine learning algorithms in terms of accuracy and sensitivity. The experimental evaluation shows that Random Forest and KNN perform better that Decision Tree and also shows that data related to the use of video lectures improves the prediction performance for some degree programs (reaching 73% in terms of accuracy and sensitivity). These preliminary results show that the approach is promising and worth exploring in future studies.
Participatory budgeting (PB) is already well established in Scotland in the form of community led grant-making yet has recently transformed from a grass-roots activity to a mainstream process or embedded 'policy instrument'. An integral part of this turn is the use of the Consul digital platform as the primary means of citizen participation. Using a mixed method approach, this ongoing research paper explores how each of the 32 local authorities that make up Scotland utilise the Consul platform to engage their citizens in the PB process and how they then make sense of citizens' contributions. In particular, we focus on whether natural language processing (NLP) tools can facilitate both citizen engagement, and the processes by which citizens' contributions are analysed and translated into policies.
Social technologies have revolutionized the world of music and playlists have become the new radios. However, the production of playlists has been mainly focused on identifying the songs that the listener might like, disregarding the songs sequencing process. In this paper, we propose a highly user-oriented sequencing method. The idea is to use the user's listening history to identify a possible sequencing criterion used by the user when playing out songs. The proposed method is then evaluated with real users and results showed the importance of personalization in the playlist sequencing process. Although the study is still in its early stage, results are promising and, in the future, we plan to personalize the sequencing process even more.
As the prevalence of algorithmic decision-making increases, so does the study of algorithmic fairness. When this aspect is disregarded, bias and discrimination are created, reproduced or amplified. Accordingly, work has been done to harmonize definitions of fairness and categorize ways to improve it. While using demographic data about the protected group is a possible solution, in real-world applications privacy concerns as well as uncertainty about the relevant attributes make it unrealistic. Consequently, we seek in this work to provide an overview of the methods that do not require such data, to identify which areas might be under-researched and to propose research questions for the first phase of the PhD. The influence of datasets size in the discovery and mitigation of unknown biases appears to be such an area, one that we plan to explore more fully during the thesis.