Papers by Antonella Carbonaro
Interpretability of AI Systems in Electronic Governance
Electronic Governance with Emerging Technologies
Sensors
Despite the pervasiveness of IoT domotic devices in the home automation landscape, their potentia... more Despite the pervasiveness of IoT domotic devices in the home automation landscape, their potential is still quite under-exploited due to the high heterogeneity and the scarce expressivity of the most commonly adopted scenario programming paradigms. The aim of this study is to show that Semantic Web technologies constitute a viable solution to tackle not only the interoperability issues, but also the overall programming complexity of modern IoT home automation scenarios. For this purpose, we developed a knowledge-based home automation system in which scenarios are the result of logical inferences over the IoT sensors data combined with formalised knowledge. In particular, we describe how the SWRL language can be employed to overcome the limitations of the well-known trigger-action paradigm. Through various experiments in three distinct scenarios, we demonstrated the feasibility of the proposed approach and its applicability in a standardised and validated context such as SAREF
Computing
The huge volume of data gathered from wearable fitness devices and wellness appliances, if effect... more The huge volume of data gathered from wearable fitness devices and wellness appliances, if effectively analysed and integrated, can be exploited to improve clinical decision making and to stimulate promising applications, as they can provide good measures of everyday patient behaviour and lifestyle. However, several obstacles currently limit the true exploitation of these opportunities. In particular, the healthcare landscape is characterised by a pervasive presence of data silos which prevent users and healthcare professionals from obtaining an overall view of the knowledge, mainly due to the lack of device interoperability and data representation format heterogeneity. This work focuses on current, important needs in self-tracked health data modelling, and summarises challenges and opportunities that will characterise the community in the upcoming years. The paper describes a virtually integrated approach using standard Web Semantic technologies and Linked Open Data to cope with he...
Recommender systems have become an important research area since the emergence of the first resea... more Recommender systems have become an important research area since the emergence of the first research paper on collaborative filtering in the mid-1990s. In general, recommender systems directly help users to select content, products, or services by aggregating and analysing historical data including suggestions from other users, and turning them into predictions of users' possible future preferences. Recommender systems combine ideas from user profiling, information filtering, data mining, machine learning and social networking to provide personalized and meaningful recommendations. For example, while standard search engines are very likely to generate the same results to the same search queries entering from different users, recommender systems are able to generate results that are personalized taking into account the individual user's profile. In general, two recommendation techniques have come to dominate: content-based filtering (CBF) and collaborative filtering (CF). In general, the first approach recommends to a user items whose content is similar to content that the user has previously viewed or selected. This has been used mainly in the context of recommending items such as books, movies, web pages, news, etc. for which informative content descriptors do exist. To accurately represent and update the features of the items is expensive, time consuming, error-prone and highly subjective. On the other hand, collaborative filtering collects information about user's rated items and makes recommendations based on items which were highly rated by users with similar profile. CF algorithms generally compute the overall similarity between users, and use that as a weight when making recommendations. Therefore, the CF techniques can be applied to virtually any kind of items and promise to scale well to large item bases becoming the most widely used approach for building online recommender systems. Finally, some systems combine both content and collaborative filtering approaches to make recommendations.
Modelling Educational Resources to Support Accessibility Requirements
2021 IEEE 18th Annual Consumer Communications & Networking Conference (CCNC)
The design and construction of personalized and flexible e-learning systems can benefit from the ... more The design and construction of personalized and flexible e-learning systems can benefit from the use of sophisticated knowledge modeling and data-analysis techniques to provide learning resources that are accessible by anyone, including learners with special accessibility needs and preferences. The paper describes a personalized educational resources system that considers learner's needs, especially when it comes to learners with accessibility requirements. We use Semantic Web technologies as resource representation formalism to provide learners with suitable learning objects, learning activities, and teaching methods based on their different preferences and needs. The system supports the creation of interactive semantic queries to bring the result sets closer to user research requirements.
Global Impacts of ICT-Qualified Worker Shortage: Exploring the Need for Educational, Firm-Based, and Societal Investments in ICT Human Capital
In today\u2019s Industry 4.0 environment, recruiting, training, and retaining highly qualified IC... more In today\u2019s Industry 4.0 environment, recruiting, training, and retaining highly qualified ICT-ready professionals remains a problem for many organizations in-cluding educational, governmental, healthcare, and business organizations. Through an in-depth literature review and syntheses, we respond to the question \u201chow can we develop ICT human capital in our global economy in an equitable, inclusive, and purposeful manner such that not organizations thrive, but also to promote social justice and equity in our global economy?\u201d In particular, the pa-per analyses the major obstacles that need to be overcome in order to direct edu-cational aspirations towards STEM education pathways, so that the demand for more people trained in ICT can be met, describing the inequalities of opportuni-ty and the transformations, both internal and external, to which higher education institutions are called to undergo. To accomplish this broad research goal, we in-tend to utilize a grounded theory approach, utilizing rich data that informs the development of frameworks that address the ICT human capital shortage as well as encouraging equitable, inclusive and purposeful methods to diversify ICT-qualified workers. Data that informs this work will be drawn from varied organ-izations, including HEI, business, government, and healthcare and education
Proceedings of the 12th International Joint Conference on Knowledge Discovery, Knowledge Engineering and Knowledge Management, 2020
The use of knowledge graphs (KGs) in advanced applications is constantly growing, as a consequenc... more The use of knowledge graphs (KGs) in advanced applications is constantly growing, as a consequence of their ability to model large collections of semantically interconnected data. The extraction of relational facts from plain text is currently one of the main approaches for the construction and expansion of KGs. In this paper, we introduce a novel unsupervised and automatic technique of KG learning from corpora of short unstructured and unlabeled texts. Our approach is unique in that it starts from raw textual data and comes to: i) identify a set of relevant domain-dependent terms; ii) extract aggregate and statistically significant semantic relationships between terms, documents and classes; iii) represent the accurate probabilistic knowledge as a KG; iv) extend and integrate the KG according to the Linked Open Data vision. The proposed solution is easily transferable to many domains and languages as long as the data are available. As a case study, we demonstrate how it is possible to automatically learn a KG representing the knowledge contained within the conversational messages shared on social networks such as Facebook by patients with rare diseases, and the impact this can have on creating resources aimed to capture the "voice of patients".
Towards Consistent Data Representation in the IoT Healthcare Landscape
Proceedings of the 2018 International Conference on Digital Health, 2018
Nowadays, the enormous volume of health and fitness data gathered from IoT wearable devices offer... more Nowadays, the enormous volume of health and fitness data gathered from IoT wearable devices offers favourable opportunities to the research community. For instance, it can be exploited using sophisticated data analysis techniques, such as automatic reasoning, to find patterns and, extract information and new knowledge in order to enhance decision-making and deliver better healthcare. However, due to the high heterogeneity of data representation formats, the IoT healthcare landscape is characterised by an ubiquitous presence of data silos which prevents users and clinicians from obtaining a consistent representation of the whole knowledge. Semantic web technologies, such as ontologies and inference rules, have been shown as a promising way for the integration and exploitation of data from heterogeneous sources. In this paper, we present a semantic data model useful to: (1) consistently represent health and fitness data from heterogeneous IoT sources; (2) integrate and exchange them; and (3) enable automatic reasoning by inference engines.
Models and services for mobile learning systems
Proceedings of the Mobile HCI, 2003
... Publication details. Download, http://giorgio.usr.dsi.unimi.it/old_site/index.html. Repositor... more ... Publication details. Download, http://giorgio.usr.dsi.unimi.it/old_site/index.html. Repository, NetMob Unitn OAI repository (Italy). Type, Inproceedings. Coverage, Milano. ...
Deploying ad-hoc learning environments to use and represent data from multiple sources and networ... more Deploying ad-hoc learning environments to use and represent data from multiple sources and networks and to dynamically respond to user demands could be very expensive and ineffective in the long run. Moreover, most of the available data is wasted without extracting potentially useful information and knowledge because of the lack of established mechanisms and standards. It is preferable to focus on data availability to choose and develop interoperability strategies suitable for smart learning systems based on open standards and allowing seamless integration of third-party data and custom applications. This paper highlights the opportunity to take advantage of emerging technologies, like the linked open data platforms and automatic reasoning to effectively handle the vast amount of information and to use data linked queries in the domain of cognitive smart learning systems.
Sensors, 2021
The automatic extraction of biomedical events from the scientific literature has drawn keen inter... more The automatic extraction of biomedical events from the scientific literature has drawn keen interest in the last several years, recognizing complex and semantically rich graphical interactions otherwise buried in texts. However, very few works revolve around learning embeddings or similarity metrics for event graphs. This gap leaves biological relations unlinked and prevents the application of machine learning techniques to promote discoveries. Taking advantage of recent deep graph kernel solutions and pre-trained language models, we propose Deep Divergence Event Graph Kernels (DDEGK), an unsupervised inductive method to map events into low-dimensional vectors, preserving their structural and semantic similarities. Unlike most other systems, DDEGK operates at a graph level and does not require task-specific labels, feature engineering, or known correspondences between nodes. To this end, our solution compares events against a small set of anchor ones, trains cross-graph attention ne...
Invited Papers Il Web fornisce sia molteplici sorgenti che contengono dati utili e rilevanti in u... more Invited Papers Il Web fornisce sia molteplici sorgenti che contengono dati utili e rilevanti in un processo di e-Learning, che informazioni non facilmente reperibili. Il Web dei Dati è un approccio basato su tecnologie semantiche verso cui diversi ricercatori si stanno muovendo per sfruttare l'interoperabilità dei repository presenti verificando che l'approccio Linked Data ha il potenziale per affrontare il problema dell'enorme disponibilità di risorse del Web così come per produrre sistemi di e-Learning personalizzati e adattivi. L'articolo presenta un sistema di estrazione automatica dei concetti usato per migliorare un ambiente di ricerca personalizzato. Tale approccio può essere considerato come una possibile istanza di un concetto più generale che consiste nella transizione dal Document Web al Document/Data Web e la conseguente gestione di questi immensi volumi di dati.
An ANT Heuristic for the Frequency Assignment Problem
The problem considered in this paper consists in defining an assignment of frequencies to radio l... more The problem considered in this paper consists in defining an assignment of frequencies to radio links, to be established between base stations and mobile transmitters, which minimizes the global interference over a given region. This problem is NP-hard and few results have been reported on techniques for solving it to optimality. We applied to this version of the frequency assignment problem an ANTS metaheuristic, that is an approach following the ant system optimization paradigm. Computational results, obtained on a number of standard problem instances, testify the effectiveness of the proposed approach.
Towards Rare Disease Knowledge Graph Learning from Social Posts of Patients
Research and Innovation Forum 2020, 2021
Support educational resource accessibility requirements in a knowledge graph modelling framework
2021 International Conference on Advanced Learning Technologies (ICALT), 2021
Knowledge modelling techniques based on knowledge graphs can be used in the context of building c... more Knowledge modelling techniques based on knowledge graphs can be used in the context of building customized and flexible e-learning systems to provide learning resources that are accessible to anyone, including learners with special needs and accessibility preferences. The paper describes how a personalized educational resource system that considers the needs of learners, especially when it comes to learners with accessibility needs, can benefit from knowledge graph-based technologies used as resource representation formalism. Knowledge graphs can be used to provide learners with suitable learning objects, learning activities, and teaching methods based on their different preferences and needs and to perform domain knowledge reasoning through modelled domain knowledge.
Personalized Views Of Relevant Information Using A Collaborative Bookmark Management System
Semantic Content Representation to Meet the Challenges of Video Analysis in Multimedia Entertainment
2018 IEEE 29th Annual International Symposium on Personal, Indoor and Mobile Radio Communications (PIMRC), 2018
This paper proposes an ontological approach for enabling semantic-aware video content annotation ... more This paper proposes an ontological approach for enabling semantic-aware video content annotation and retrieval in order to facilitate content video analysis. Semantic Web ontologies express key entities and relationships which describe videos in a formal machine-processable representation. Ontology-based knowledge representation systems can be used for content analysis, concept recognition, reasoning processes and for enabling user-friendly and intelligent multimedia content retrieval, thus meeting the challenges of video analysis.
A framework for knowledge acquisition in intelligent tutoring systems
This paper discusses a general framework for knowledge acquisition and management in an intellige... more This paper discusses a general framework for knowledge acquisition and management in an intelligent tutoring system. The system is based on constrained-based tutor paradigm in which the idea of “learning by performance error” is exploited. Such a theory states that in a given domain of knowledge there is a set of constraints that must be satisfied in order to provide the correct solution to the problem. In our particular case constraints are represented in the form of conditional statements (and in general as first order formula) and we aim at an easy mechanism for constraint acquisition. We propose a framework for acquiring and manage knowledge in an intelligent tutoring system and discuss the part related to the specific task of knowledge acquisition.
Designing Models and Services for Learning Management Systems in Mobile Settings
Lecture Notes in Computer Science, 2004
The paper presents the guidelines of a project of three Italian Universities (Bologna, Siena, Tre... more The paper presents the guidelines of a project of three Italian Universities (Bologna, Siena, Trento) which aim is to investigate the use of mobile computing technologies to support the learning processes in a University context. The project covers three main areas. The first area is concerned with finding effective models for mobile learning. The second regards the evaluation of learning processes in mobile learning environments. The third focuses on the technological aspects of mobile learning, and on their integration with e-Learning systems ...
Uploads
Papers by Antonella Carbonaro