Centro di Ricerca, Sviluppo e Trasferimento tecnologico
Social Thingum è un Centro di Trasferimento Tecnologico (CTT) che svolge attività di formazione, accompagnamento all’innovazione e trasferimento tecnologico, oltre a fornire servizi specialistici a favore delle imprese operanti in specifici ambiti economici e industriali individuati dal Ministero dello Sviluppo Economico.
Si tratta di società ed enti che dimostrano di realizzare progetti tecnologici in settori come la manifattura additiva, la realtà aumentata, l’internet delle cose, il cloud, la cybersicurezza e l’analisi dei big data.
Il certificato viene rilasciato attraverso un sistema camerale da Unioncamere che svolge un ruolo importante nel processo di digitalizzazione delle imprese e, più in generale, di modernizzazione del Paese attraverso la realizzazione di progetti e l’erogazione di servizi.
Social Thingum è uno dei 28 centri di trasferimento tecnologico in Italia, la cui specializzazione è stata riconosciuta negli ambiti dell’Industrial Internet, Internet of Things e/o Internet of Machines, Big Data e Analytics, legate alle attività di:
- Formazione tecnologica
- Formazione trasferimento tecnologico su modelli organizzativi e/o di business, tecnica e industriale
- Progettazione e pianificazione interventi di implementazione di tecnologie Industria 4.0
- Ricerca industriale e sviluppo sperimentale inclusa prototipazione anche virtuale
- Produzione di materiale tecnico o documentale sulle tecnologie 4.0
Social Thingum è un anche Centro di Ricerca, Sviluppo e Trasferimento tecnologico accreditato dalla Regione Lombardia mediante Questio.
In questa veste, possiamo aiutare le aziende ad ottenere finanziamenti a fondo perduto o agevolazioni finanziarie per la realizzazione di progetti di Ricerca e Sviluppo (R&D) e/o di innovazione di prodotto e/o processo. Queste opportunità di finanziamento sono concrete e si realizzeranno anche a breve. Infatti, sono stati pubblicati due bandi di finanziamento per iniziative di Ricerca ed Innovazione: il bando per Innovation Manager e il bando Innodriver.
Ci proponiamo di assistervi nella stesura della documentazione necessaria alla partecipazione al bando e, soprattutto, nella realizzazione dei progetti stessi. Infatti, essendo Social Thingum anche Centro di Ricerca accreditato, i nostri costi sono spesabili all'interno del progetto.
In particolare, abbiamo anche ben sette Innovation Manager accreditati dal MISE, in grado di offrire un'ampia varietà di servizi a supporto di progetti di innovazione per le aziende.
Social Thingum è partner di AI Magister uno dei Poli Europei di innovazione digitale (EDIH), sportelli unici che aiutano le imprese e le Pubbliche Amministrazioni a migliorare i processi di business e di produzione, i prodotti o i servizi che utilizzano tecnologie digitali.
AI Magister sostiene la Digital & Intelligent Green Transformation in settori chiave, rendendo la tecnologia accessibile e vantaggiosa per un’ampia gamma di utenti con l’obiettivo di:
- Favorire la crescita di molteplici organizzazioni: collaborando con start-up, PMI e grandi imprese per integrare efficacemente l’IA nei loro processi operativi.
- Migliorare l’efficienza operativa: integrando l’Intelligenza Artificiale nei processi fondamentali, aiutiamo le organizzazioni a ottimizzare le loro operazioni per una maggiore efficienza e sostenibilità.
- Aumentare competenze e conoscenze aziendali: Migliorando le condizioni attraverso upskilling e reskilling e diminuendo il gap formativo
Social Thingum è una delle 100 realtà di intelligenza artificiale più innovative in Europa, certificata da European Digital SME Alliance.
La European DIGITAL SME Alliance è la più grande rete di piccole e medie imprese ICT in Europa, che rappresenta più di 45.000 imprese in totale. Rappresenta le associazioni nazionali delle PMI degli stati membri dell'UE presso il Joint Research Center della European Commission, con l'obiettivo di mettere la PMI digitale al centro dell'agenda dell'UE. La European DIGITAL SME Alliance:
- Partecipa e propone progetti finanziati dall'UE;
- organizza conferenze e seminari;
- produce newsletter e studi;
- svolge attività di ricerca e sviluppo su aree correlate all'ICT;
- monitora le politiche e i regolamenti dell'UE sull'ICT e informa i suoi membri.
Pubblicazioni
Autori: Stefano Valtolina, Ricardo Anibal Matamoros Aragon, Francesco Epifania
Tag: Conversational interface, Machine learning for education, End User Development objects, Acceptability and Usability
This paper presents a software framework that enables teachers to design reliable, personalised conversational agents tailored to their pedagogical goals and student learning preferences. The system combines a Retrieval-Augmented Generation (RAG) architecture with a visual configuration environment, allowing educators to upload, validate, and organise domain-specific teaching materials into a teacher-curated content corpus. Educators can configure adaptive tutoring strategies based on the VARK model (Visual, Auditory, Reading/Writing, Kinesthetic), allowing the conversational agents to address diverse learning preferences and educational contexts. Unlike fully autonomous or black-box educational AI systems, this approach foregrounds teacher agency and pedagogical alignment, enabling intuitive control over content and interaction style. A preliminary evaluation with university educators assessed usability (SUS), perceived utility (UTAUT), cognitive load (NASA-TLX), and creative-technical capacity (CTS), revealing promising results and informing future design directions. The system supports the development of human-centred AI tutors that are transparent, con figurable, and grounded in teacher expertise.
Autori: Stefano Valtolina, Ricardo Anibal Matamoros, Francesco Epifania, Alessia Orlandi
This paper presents an AI-driven conversational agent to support self-directed learning through a personalised pedagogical framework. Leveraging the Visual, Aural, Read/Write, and Kinesthetic (VARK) model, the system adapts educational content to learners’ individual preferences. Retrieval-Augmented Generation (RAG) enhances response accuracy and minimises hallucinations commonly associated with Large Language Models (LLMs). A user study compared the chatbot's performance to a traditional search engine in engagement, usability, and learning outcomes. The search engine was selected as a baseline to contrast conventional keyword-driven information retrieval with adaptive, dialogue-based guidance. Results indicate that personalised, AI-supported interactions foster greater engagement and knowledge retention. The study highlights the potential and limitations of VARK-driven AI tools in personalised learning.
Autori: Stefano Valtolina, Ricardo Anibal Matamoros, Francesco Epifania
Tag: Conversational interface, Machine learning for education, Learning objects
In recent years, we have seen a significant proliferation of e-learning platforms. E-learning platforms allow teachers to create digital courses in a more effective and time-saving way, but several flaws hinder their actual success. One main problem is that teachers have difficulties finding and combining open-access learning materials that match their specific needs precisely when there are so many to choose from. This paper proposes a new strategy for creating digital courses that use learning objects (LOs) as primary elements. The idea consists of using an intelligent chatbot to assist teachers in their activities. Defined using RASA technology, the chatbot asks for information about the course the teacher has to create based on her/his profile and needs. It suggests the best LOs and how to combine them according to their prerequisites and outcomes. A chatbot-based recommendation system provides suggestions through BERT, a machine-learning model based on Transformers, to define the semantic similarity between the entered data and the LOs metadata. In addition, the chatbot also suggests how to combine the LOs into a final learning path. Finally, the paper presents some preliminary results about tests carried out by teachers in creating their digital courses.
Autori: Sahar Shah, Sara Lucia Manzoni, Farooq Zaman, Fatima Es Sabery, Francesco Epifania, Italo Francesco Zoppis
Tag: Continual learning, natural language processing, text classification, fine-tuning, Distil-BERT
Continual learning (CL) with bidirectional encoder representation from transformer (BERT) and its variant Distil-BERT, have shown remarkable performance in various natural language processing (NLP) tasks, such as text classification (TC). However, the model degrading factors like catastrophic forgetting (CF), accuracy, task dependent architecture ruined its popularity for complex and intelligent tasks. This research article proposes an innovative approach to address the challenges of CL in TC tasks. The objectives are to enable the model to learn continuously without forgetting previously acquired knowledge and perfectly avoid CF. To achieve this, a task-independent model architecture is introduced, allowing training of multiple tasks on the same model, thereby improving overall performance in CL scenarios. The framework incorporates two auxiliary tasks, namely next sentence prediction and task identifier prediction,to capture both the task-generic and task-specific contextual information. The Distil-BERT model, enhanced with two linear layers, categorizes the output representation into a task-generic space and a task-specific space. The proposed methodology is evaluated on diverse sets of TC tasks, including Yahoo, Yelp, Amazon, DB-Pedia, and AG-News. The experimental results demonstrate impressive performance across multiple tasks in terms of F1 score, model accuracy, model evaluation loss, learning rate, and training loss of the model. For the Yahoo task, the proposed model achieved an F1 score of 96.84 %, accuracy of 95.85 %, evaluation loss of 0.06, learning rate of 0.00003144. IntheYelptask,ourmodelachievedanF1 score of 96.66%, accuracy of 97.66 %, evaluation loss of 0.06, and similarly minimized training losses by achieving the learning rate of 0.00003189. For the Amazon task, the F1 score was 95.82 %, the observed accuracy is 97.83%, evaluation loss was 0.06, and training losses were effectively minimized by securing the learning rate of 0.00003144. In the DB-Pedia task, we achieved an F1 score of 96.20 %, accuracy of 95.21 %, evaluation loss of 0.08, with learning rate 0.0001972 and rapidly minimized training losses due to the limited number of epochs and instances. In the AG-News task, our model obtained an F1 score of 94.78 %, accuracy of 92.76%, evaluation loss of 0.06, and fixed the learning rate to 0.0001511. These results highlight the exceptional performance of our model in various TC tasks, with gradual reduction in training losses over time, indicating effective learning and retention of knowledge.
Autori: Stefano Valtolina, Ricardo Anibal Matamoros, Francesco Epifania
Recently, there has been an increasing interest in language-based interactions with technology. Driven by the success of intelligent personal assistants, the number of conversation-based interactions is growing in several domains. Nevertheless, the literature highlights a lack of models specifically studied to analyse communication blocks and the level of accessibility and acceptability by the user that can characterise Human-Agent dialogue. This paper aims to learn how much an agent can be accessible, how much the communication is understandable, and if it brings to a successful conclusion by extending two known models, the UTAUT2 and CEM. For both models, we defined new indicators to analyse communicability, acceptability degrees and accessibility level using WCAG guidelines. We designed two conversational agents to test our models and conducted preliminary tests. A first agent is used to assist students in following remote digital courses, and a second to help older people in their daily activities and to monitor the indexes of active life defined to control the trend of older people’s physical-cognitive state.
Autori: Matteo Garavaglia, Alessandro Solinas, Ricardo Anibal Matamoros Aragon, Stefania Bandini, Francesco Epifania
Tag: Recommender Systems, Impressions, Cold Start, Real-time Recommendations
Recommender Systems (RS) are tools that are often utilized and need constant development in both the structure and the data to use. Impressions Data are a new type of information that is underused and can be helpful in various scenarios. Therefore, we propose a hybrid RS that uses Impressions to mitigate the significant issues in our original system, Knowledge Graph Attention Network (KGAT). The first problem is the situation of complete cold-start, for which we propose the use of questions on selected meaningful attributes and a BERT-based Content-Based RS to perform recommendations following the user’s choices. After that, when in a framework of semi cold-start, the recommendations will be enhanced by using Impressions to rerank the following ones and, from these interactions, to build a profile to use with KGAT. The last issue we will address is the need for more interpretation of negative interactions through the Knowledge Graph, that is, recommendations presented but not chosen. To solve this issue, we use the Impression Discounting model on the set of recommendations produced by KGAT.
Autori: Alberto Schiaffino, Matteo Reina, Ricardo Anibal Matamoros Aragon, Alessandro Solinas, Francesco Epifania
Tag: Anomaly detection, Content Management Systems, Cybersecurity
In cybersecurity, protecting IT infrastructures from potential threats is one of the most pressing and intricate challenges. Over recent years, we have witnessed an exponential surge in techniques aimed at enhancing defense processes against attacks orchestrated by malicious users. Zero-day attacks are among the most insidious and damaging types of attacks. A Zero-Day attack exploits software vulnerabilities that are unknown to manufacturers or the general public. Due to their concealed nature, these vulnerabilities pose a particularly severe threat: anyone who becomes aware of them before the manufacturers can exploit them, launching attacks with a high likelihood of success since there are no existing solutions or patches to counteract them. Given the importance of preventing and detecting such attacks, we propose an innovative approach based on Deep Learning techniques in this paper. The goal is to conduct anomaly detection analysis on system logs of a Content Management System (CMS) platform. To achieve this, we utilized an algorithm known as DeepLog, which leverages Long Short-Term Memory (LSTM) neural networks to identify patterns in the sequential nature of the logs.
Autori: Simone Re, Matteo Olivieri, Ricardo Anibal Matamoros Aragon, Alessandro Solinas, Francesco Epifania
Tag: Anomaly Detection, Artificial Intelligence, E-learning, Attention Mechanism
In today’s interconnected digital landscape, the Internet plays a pivotal role in our daily human activities. However, the intricacy of the online communication network exposes vulnerabilities that can be exploited by malicious actors, who adopt increasingly sophisticated strategies to compromise cybersecurity. This issue extends to the domain of e-learning, where the protection of user personal data and the interaction with external educational resources become critical aspects. In this context, we introduce an e-learning platform developed by Informattiva, integrated with an advanced cybersecurity mechanism. This mechanism is designed to analyze educational resources from external repositories, such as Merlot. org, aiming to identify potential insecurities based on URLs. To achieve this, we implemented a model based on the Bidirectional Gated Recurrent Unit (BiGRU) with attention mechanisms, focusing on the identification of potentially malicious web addresses. Preliminary results indicate that, through bidirectional processing and attention mechanisms, our methodology has the potential to effectively differentiate suspicious URLs from secure ones.
Autori: Danish Ali, Francesco Epifania, Naeem Ahmed, Bilal Khan, Luca Marconi, Ricardo Matamoros
Tag: Convolutional Neural Network, Recognition System, Gender Prediction, Age Prediction
Age and gender information are essential for many real-world applications, such as social intelligence, biometric identity verification, video surveillance, human-computer interaction, digital consumer, crowd behavior analysis, online marketing, item recommendation, and many more. This study intends to employ deep learning technology in the prediction process, effective accuracy, and predictive mining and assess it in order to obtain the best outcomes of prediction and get around the issues of time, accuracy, and processing load. In this multi-task learning problem, age and gender are predicted concurrently with the help of a single Convolutional neural network with two heads (output branches). The model has 95% accuracy for gender classifier and 92% accuracy for age classifier. The pro-posed model uses the computing resources (RAM, CPU, and GPU) in a much more optimized manner and the computing cost is also lower.
Autori: Luca Marconi, Ricardo A. Matamoros A, Francesco Epifania
Tag: Machine Learning, Recommender Systems, Artificial Intelligence, eXplainable Artificial Intelligence, eXplainable Recommender Systems
The use of learning objects (LOs) to create digital courses has been widely advocated by learning strategists and by teachers engaged in the e-learning domain. The ability to combine chunks of learning material as to meet complex educational requirements is still a challenge. This paper explores the idea that a learning assistant advises teachers about the e-learning modules to take into account for their courses. An AI-based digital assistant can provide significant opportunities, but might be perceived as a threat. The paper presents how teacher could perceive a virtual assistant as more trustworthy when it applies interactive visual strategies. To analyze teachers’ acceptance of the digital assistant, our proposal aims at extending the Unified Theory of Acceptance and Use of Technology (UTAUT) model in order to incorporate three new constructors: Communicability, perceived trust and experience. To this end, 14 teachers have been involved in a user tests.
Autori: Stefano Valtolina, Ricardo Anibal Matamoros, Elia Musiu, Francesco Epifania, Mattia Villa
TagHuman Computer Interaction, Recommender System, UTAUT, Interactive Visualization
The use of learning objects (LOs) to create digital courses has been widely advocated by learning strategists and by teachers engaged in the e-learning domain. The ability to combine chunks of learning material as to meet complex educational requirements is still a challenge. This paper explores the idea that a learning assistant advises teachers about the e-learning modules to take into account for their courses. An AI-based digital assistant can provide significant opportunities, but might be perceived as a threat. The paper presents how teacher could perceive a virtual assistant as more trustworthy when it applies interactive visual strategies. To analyze teachers’ acceptance of the digital assistant, our proposal aims at extending the Unified Theory of Acceptance and Use of Technology (UTAUT) model in order to incorporate three new constructors: Communicability, perceived trust and experience. To this end, 14 teachers have been involved in a user tests.
CMS Optimisation with Deep Learning Techniques (2021)
Autori: Alberto Schiaffino, Matteo Reina, Ricardo A. Matamoros A., Francesco Epifania, Francesco Ruggeri, Ignazio Maria Castrignano, Luca Marconi
Tag: Machine Learning · Content Management Systems · Artificial Intelligence · Sentiment Analysis · Natural Language Processing
This study stems from the observation of the growing demand, interest and incredible potential offered by Artificial Intelligence and Machine Learning techniques in the fields of content website or ecommerce. In particular, we will study the Content Management System (CMS), a system capable of creating and maintaining websites, forums and applications for customers. These systems have an intrinsic predisposition to provide services based on Machine Learning (ML), thus being able to add to the solutions offered by the most innovative ML systems. These include Sentiment Analysis, the Recommender System and the more innovative Chatbot . Integrating them into a website or web platform very often leads to significantly improved results for the user-experience offered to the user. This study aims to analyse the new artificial intelligence techniques and how to implement them in a company. Reporting also a case of Study based on a real agency.
Autori: Antonio Crinieri, Luca Terzi, Francesco Ruggeri, Ricardo A. Matamoros A., Francesco Epifania, Luca Marconi
Tag: Transfer Learning · Skin Diseases · Convolutional Neural Networks
This study deals with the problem of the recognition of dermatological and exanthematic diseases through the use of deep learning techniques in order to diagnose malignant diseases at an early stage and in general to bring the pathology identified by the models to the attention of the person. A fundamental part of the research was the study of the methodologies present in the state of the art and for this reason in this paper we report the studies considered as most relevant. In this paper, two different types of models are reported, the Convolutional Network Disease (CND) model and the CND-InceptionV3 model using the transfer learning technique. The use of these two models made it possible to carry out an experimental phase in which the performance that can be achieved using the ISIC-archive dataset was analysed. Subsequently, the description of the work carried out for the improvement of the dataset through the association of syntactic-semantic information is reported. Finally, in the last section, conclusions are drawn on the values obtained and future developments that can be made to improve the performance of the models reported are reported.
Autori: Maurizio Monticelli, Ricardo A. Matamoros A., Francesco Epifania, Luca Marconi, Antonio De Simone
Tag: Recommender Systems · Clustering · E-commerce · Machine Learning · Software Components
In recent years, the exponential growth in the number of items and products handled by e-commerce sites has led to the introduction of intelligent systems aimed at supporting users during the decisionmaking proces. Making the choice of a product among thousands of items becomes complicated for consumers, and in response to this problem, recommender systems (RS) are born. These systems are a set of algorithms based on the concept of information filtering and make it possible to reduce the cognitive effort required of users. In this paper we present a model-based RS, belonging to the collaborative filtering (CF) category, for the e-commerce website of the company Nathan Instruments (NI). Thus, the main objective of this paper is to provide an intelligent approach for recommending configurations of hardware components for Computers. This configurator uses clustering algorithms to address the problems associated with small dataset sizes. Finally, in the experimentation and conclusion sections it is reported how the proposed model simplifies the decision process related to the required computer customization in terms of hardware and software components
Autori: Luca Marconi, Ricardo Anibal Matamoros Aragon, Italo Zoppis, Sara Manzoni, Giancarlo Mauri, Francesco Epifania
Tag: Explainable AI, Personalized learning, WhoTeach, Social recommendations, Graph attention networks
Learning and training processes are starting to be affected by the diffusion of Artificial Intelligence (AI) techniques and methods. AI can be variously exploited for supporting education, though especially deep learning (DL) models are normally suffering from some degree of opacity and lack of interpretability. Explainable AI (XAI) is aimed at creating a set of new AI techniques able to improve their output or decisions with more transparency and interpretability. In the educational field it could be particularly significant and challenging to understand the reasons behind models outcomes, especially when it comes to suggestions to create, manage or evaluate courses or didactic resources. Deep attentional mechanisms proved to be particularly effective for identifying relevant communities and relationships in any given input network that can be exploited with the aim of improving useful information to interpret the suggested decision process. In this paper we provide the first stages of our ongoing research project, aimed at significantly empowering the recommender system of the educational platform “WhoTeach” by means of explainability, to help teachers or experts to create and manage high-quality courses for personalized learning. The presented model is actually our first tentative to start to include explainability in the system. As shown, the model has strong potentialities to provide relevant recommendations. Moreover, it allows the possibility to implement effective techniques to completely reach explainability.
Autori: Marconi, L., Matamoros Aragon, R., Zoppis, I., Manzoni, S., Mauri, G., & Epifania, F.
Tag: Graph Attention Networks; Social Networks; Social Recommendations; WhoTeach;
Learning and training processes are starting to be affected by the diffusion of Artificial Intelligence (AI) techniques and methods. AI can be variously exploited for supporting education, though especially deep learning (DL) models are normally suffering from some degree of opacity and lack of interpretability. Explainable AI (XAI) is aimed at creating a set of new AI techniques able to improve their output or decisions with more transparency and interpretability. Deep attentional mechanisms proved to be particularly effective for identifying relevant communities and relationships in any given input network that can be exploited with the aim of improving useful information to interpret the suggested decision process. In this paper we provide the first stages of our ongoing research project, aimed at significantly empowering the recommender system of the educational platform”WhoTeach” by means of explainability, to help teachers or experts to create and manage high-quality courses for personalized learning. The presented model is actually our first tentative to start to include explainability in the system. As shown, the model has strong potentialities to provide relevant recommendations. Moreover, it allows the possibility to implement effective techniques to completely reach explainability
Autori: Luca Marconi, Ricardo Anibal Matamoros Aragon, Serena Fossati, Italo Zoppis, Rossana Actis Grosso, Sara Manzoni, Giancarlo Mauri,and Francesco Epifania
Tag: Social Networks, WhoTeach, Social Recommendations, Graph Attention Networks
Nowadays, learning and training processes are beginning to be affected by the diffusion of Artificial Intelligence (AI) techniques and methods. Despite its potentialities, AI and in particular deep learning (DL) models are normally suffering from some degree of opacity and lack of interpretability. Explainable AI (XAI) is aimed at creating a set of new AI techniques able to improve their output or decisions with more transparency and interpretability. Among these techniques, deep attentional mechanisms provide the possibility to improve the performances of the output of the models and especially they allow to explain the reason why a specific output is given. In this paper we describe the current stage of explainability for a collaborative-filtering recommender system (RS) of the “WhoTeach” educational platform, which is aimed at designing of new didactic programs and courses, particularly by the means of the results of an experimentation that has been performed on a selected set of users, by means of the cooperative evaluation approach. A relevant result of the experimentation shows the need for improving explainability and suggested a development direction towards attentional mechanisms for WhoTeach functionalities that are devoted to suggest educational resources according to user needs and profile. This type of models allows to justify the chosen recommendations provided by the model by means of attention weights, which have also been statistically tested.
Autori: Italo Zoppis, Sara Manzoni, Giancarlo Mauri, Ricardo Aragon, Luca Marconi and Francesco Epifania
Tag: Social Networks, WhoTeach, Social Recommendations, Graph Attention Networks
Recent studies in the context of machine learning have shown the effectiveness of deep attentional mechanisms for identifying important communities and relationships within a given input network. These studies can be effectively applied in those contexts where capturing specific dependencies, while downloading useless content, is essential to take decisions and provide accurate inference. This is the case, for example, of current recommender systems that exploit social information as a clever source of recommendations and / or explanations. In this paper we extend the social engine of our educational platform “WhoTeach” to leverage social information for educational services. In particular, we report our work in progress for providing “WhoTeach” with an attentional-based recommander system oriented to the design of programmes and courses for new teachers.
Autori: Epifania F, Marconi L, Mauri G, Manzoni S, Dondi R, Zoppis I
Tag: Recommender Systems
Recommender Systems have became extremely appealing for all technology enhanced learning researches aimed to design, develop and test technical innovations which support and enhance learning and teaching practices of both individuals and organizations. In this scenario a new emerging paradigm of explainable Recommander Systems leverages social friend information to provide (social) explanations in order to supply users with his/her friends’ public interests as explained recommendation. In this paper we introduce our educational platform called “WhoTeach”, an innovative and original system to integrate knowledge discovery, social networks analysis, and educational services. In particular, we report here our work in progress for providing “WhoTeach” environment with optimized Social Explainable Recommandations oriented to design new teachers’ programmes and courses.
Autori: Andri PERL, Sebastiaan BONGERS, Simone Bassis, Bruno Apolloni, Swiss Reinsurance Co Ltd
Tag: /
An electronic, real-time system performs maneuver recognition of vehicles based on dynamically measured telematics data, particularly the sensory data of smartphone sensors, and more particularly data from the accelerometer sensor and the global positioning system (GPS) sensor and/or the gyroscope sensor of a smartphone. The axes of the smartphone may be moving independently relative to the axes of the vehicle and thus do not need to be aligned with the axes of the vehicle. Driver behaviors and operational parameters are automatically measured and discriminated, based on automatically individuated and measured driver maneuvers within various measured to vehicle trajectories, and an output signal is generated based upon derived risk measure parameters and/or crash attitude measure parameters. The system can use score-driven, especially risk-score driven, operations associated with motor vehicles or transportation modes for passengers or goods, and reliant on a dynamic, telematics-based data aggregation and dynamically measured driving maneuvers, respectively.
Autori: Bruno Apolloni, Simone Bassis
Tag: Algorithmic inference machine learning parameter distribution confidence intervals learning Boolean functions
We start from the very operational perspective – having data, organize them in a suitable way to be used in the future – to enter the long standing fray on the nature of inferred parameters within a machine learning thread. Still in an operational perspective, we introduce a parametric inference approach that unprecedentedly gets rid of most drawbacks incurred by current methods to compute confidence intervals. The key idea is to consider the parameters of the distribution underlying a sample to be random, where randomness is expressed in terms of a probability measure of the compatibility of the parameter values with the actually observed data. The probability is understood, in a frequentist acceptation, in terms of the asymptotic frequency of those parameter values matching the observed sample in a story of infinite observations. The aim of this paper is to recap and complete theoretical results obtained through our approach as presented in preceding papers. In particular, here we focus on statistical tools both for computing confidence regions, at the basis of appraising the learnability of a function, and for checking their efficacy. We basically support our theory with a series of well-known benchmarks where, as for both volume and coverage of the confidence regions, our method proves superior – with very few ties – to those of competitors. Then we mention some results in computational learning theory that have been achieved recently exactly by adopting our approach, with a special focus on a new data_ accuracy - sample_complexity trade off.
Autori: Bruno Apolloni, Simone Bassis, Marco Mesiti, Stefano Valtolina, Francesco Epifania
Tag: Recommender system, Decision trees, Genomic features
We introduce a new recommending paradigm based on the genomic features of the candidate objects. The system is based on the tree structure of the object metadata which we convert in acceptance rules, leaving the user the discretion of selecting the most convincing rules for her/his scope. We framed the deriving recommendation system on a content management platform within the scope of the European Project NETT and tested it on the Entree UCI benchmark.
Autori: Francesco Epifania
Tag: Social Network, LMS, RS, entrepreneurship
With the ambition of providing teachers with a novel concrete tool called “Social Intelligent Learning Management System (SILMS)” for worldwide exploiting didactic contents to feature their courses, I faced the problem of creating a social platform with adequate functionalities to satisfy the teacher expectations. This goal involved many disciplines and practices ranging from DB management, content management, social networking, till the exploitation of new cognitive systems in the thread of WEB4. 0 services. At the same time my approach was much oriented to realize a real tool of concrete usage, still with distinguishably advanced features. Thus, starting with a well designed architecture I endowed it with key functionalities that become the stakeholders of the emerging social networks: 1) a quality system ensuring the value of the materials the users put in the platform repository as their contribution to the social business, 2) a recommender system based on either ontology assisted navigator or computational intelligence techniques constituting the principal tool to guide teachers along the assembling of materials into courses.
Autori: Francesco Epifania, Riccardo Porrini
Tag: Recommender System, Learning Resources, Social Network, e-Learning, User-centric Evaluation.
The NETT Recommender System (NETT-RS) is a constraint-based recommender system that recommends learning resources to teachers who want to design courses. As for many state-of-the-art constraint-based recommender systems, the NETT-RS bases its recommendation process on the collection of requirements to which items must adhere in order to be recommended. In this paper we study the effects of two different requirement collection strategies on the perceived overall recommendation quality of the NETT-RS. In the first strategy users are not allowed to refine and change the requirements once chosen, while in the second strategy the system allows the users to modify the requirements (we refer to this strategy as backtracking). We run the study following the well established ResQue methodology for user-centric evaluation of RS. Our experimental results indicate that backtracking has a strong positive impact on the perceived recommendation quality of the NETT-RS.
Autori: B Apolloni, F Epifania, M Mesiti, M Mesenzani, S Valtolina
E’un dato singolare che nonostante le ripetute note degli organi di governo e delle imprese che rilevano la mancanza di spirito imprenditoriale e d’altro canto le benemerenze rispetto al paese di chi fa impresa, a tutt’oggi l’Educazione di Imprenditorialità non fa parte dei programmi ministeriali delle scuole di qualsiasi ordine e grado, ad eccezione di alcune Università. Non è così nel resto di Europa, come mostra la mappa in Fig. 1 (Eurydice, 2012). Benché datata di 5 anni questa mappa illustra appunto l’assenza di strategie nel settore in Italia. Questa situazione peraltro non è isolata come sottolineato dalla Comunità Europea che ha promosso varie iniziative per migliorarla (Europe, 2020). Tra queste è il progetto NETT (Networked Entrepreneurship Training of Teachers--http://www. nett-project. eu/), finanziato a un consorzio di 4 partner (due Italiani, uno Bulgaro e uno Turco) per creare una social network di educatori di impresa attraverso la realizzazione di una piattaforma ad hoc. L’idea di base è che, trattandosi di una disciplina ancora acerba, per la sua maturazione occorra il contributo di coloro che già la insegnano o prevedano di farlo nell’immediato futuro. Per questo motivo occorrono strumenti che permettano agli operatori del settore di scambiarsi materiali e di favorire la discussione e la valutazione dei materiali al fine di identificare quelli utili e di qualità per l’insegnamento di questa disciplina.
Autori: BRUNO APOLLONI, LUCA MARCONI, FRANCESCO EPIFANIA, ALESSIO ANGHILERI, MARCO MESITI, STEFANO VALTOLINA, SERENA DI GAETANO, ALBERTO SCHIAFFINO, MATTEO REINA, ROBERTO PELLEGRINI
Tag: Social appliances; green social network; machine learning; collective intelligence
We discuss a Cloud-based Collective Intelligence model and its in-progress implementation to direct users toward an optimal usage of their home appliances as a way of getting both personal advantage and an overall reduction of pollution and energy consumption. In this model sustainability is considered with respect to two types of resources: natural ones, to be mostly preserved, as indicated above, and brain resources, in terms of intention and knowledge, to be convoyed to a common target. Having the first aspect for a given, in this paper we focus on the second by examining three distinct factors: user experience, knowledge achievement and business model. Our service paradigm is rooted on a Social Network of Facts that requires experts’ know-how, like that owned by the appliance manufacturer, but exploits it in an autonomous way so as to comply with the specific intentions of the individual users. While cloud architectural and communication aspects are solved in a standard, though advanced, way, the interplay between user and experts is considered variously within a range of business models. As the success of these models is related to the network population, here we discuss some preliminary simulations based on an effectively implemented infrastructure and on the extrapolation of early collected data.
Altre Pubblicazioni dei co-founders
Prof. Francesco Epifania Ph.D: https://scholar.google.it/citations?hl=en&user=pxgD96gAAAAJ&view_op=list_works&sortby=pubdate
Prof. Bruno Apolloni: https://dblp.uni-trier.de/pers/hy/a/Apolloni:Bruno
Prof. Nikolov Roumen Ph.D: https://dblp.uni-trier.de/pers/hd/n/Nikolov:Roumen
Prof.ssa Eugenia Kovatcheva Ph.D: https://dblp.uni-trier.de/pers/hd/k/Kovatcheva:Eugenia
Simone Bassis Ph.D: https://dblp.org/pers/hd/b/Bassis:Simone