Slides for the PHD course at the DINFO dept of the University of Florence, 2014
Social Networking and knowledge, Semantic and Social Networks, Recommendations and Suggestions, Natural Language Processing System, Knowledge Representation System, Reasoning System, Sistema OSIM, Smart Cities, Open Data, LOD, Linked Open Graph, Data Mining, data modeling, knowledge management, reasoning, smart city
Knowledge Management and Natural Language Processing: OSIM, CoSkoSAM
Content and Protection Management, grid computing: AXMEDIS AXCP
Social Media, recommendations and tool: ECLAP.eu, MyStoryPlayer, Social Graph, IPR Wizard…
PROJECTS: sii-mobility, OSIM, SACVAR, ECLAP, Coll@bora.
Modelli Semantici e Gestione della Conoscenza: Social Network vs Knowledge Ma...Paolo Nesi
Modelli Semantici e Gestione della Conoscenza: Social Network vs Knowledge Management Systems. Seminario per il dottorato in Ingegneria Informatica e delle telecomunicazioni, Univ. Studi di Firenze, 2013
Sistemi Distribuiti part 5: P2P systems: from simple to distributed P2P trust...Paolo Nesi
P2P, DHT, distributed trust and certification information for DRM, P2P Health care record protection, AXMEDIS P2P, P2P for video streaming, DHT models, P2P for CDN, P2P and AXCP, monitoring P2P networks.
Smart City Strategic Forecast, SmartCity360, BratislavaPaolo Nesi
Smart City strategy, city smartening, big data amanagement,
-Taking into account results of participatory actions
-Smart city strategic plan, city agenda: prioritizing interventions
-Agreements for collaborations with main actors:
main research centers, main City Operators, etc.
-Direct collaborations on specific projects on:
ICT, Mobility, Culture, Energy, etc.
Experimenting on specific projects of the Smart City Strategic Plan
-Needs of harmonizing results and aggregating data towards dashboards
Smart Cloud Engine and Solution based on Knowledge BasePaolo Nesi
Complexity of cloud infrastructures needs models and tools for process management, configuration, scaling, elastic computing and healthiness control. This paper presents a Smart Cloud solution based on a Knowledge Base, KB, with the aim of modeling cloud resources, Service Level Agreements and their evolution, and enabling the reasoning on structures by implementing strategies of efficient smart cloud management and intelligence. The solution proposed provides formal verification tools and intelligence for cloud control. It can be easily integrated with any cloud configuration manager, cloud orchestrator, and monitoring tool, since the connections with these tools are performed by using REST calls and XML files. It has been validated in the large ICARO Cloud project with a national cloud service provider.
Km4City White Paper: Production tools for Smart City, from data to services f...Paolo Nesi
In the Smart City context, hundreds of data sets are available. Many of them are open data, accessible from local, regional, national, European public administrations, national institute of statistics, etc. These data can be static, statistical data or real time. In addition, many other data are produced by other institutions like Europeana, ECLAP, Getty, Voc, dbPedia, etc. Typically most of the data are geolocated and can be accessed as files in various formats (CSV, XLS, KMZ, JSON, XML, HTML, MySQL, ZIP, LSMA, SHP, etc.), other are accessible as Linked Data, Linked Open Data or via RDF Store end points (see dpPedia, Europeana, Senate of the Republic, Chamber of Deputies, ECLAP, km4city, etc.). Personal, private and critical data can be added to these open data. Some private data are produced by companies, like for example the position of car sharing vehicles, the position of taxis, busses, flows in the city, energy consumption data in a neighborhood, etc. Many of these data can be useful for public administrations to take decisions and to provide services. Personal data are related to a person, include personal identifications, the position of the person, its profile, etc., and need to be managed in accordance with terms of use and privacy policies. Finally, critical and personal data may be used by bad-intentioned to take actions against citizens security and infrastructures, and thus licensing and conditional access solutions are adopted.
Data are typically produced by central data producers, and many of these can provide their data in different ways and formats. Among them: traffic management systems, fleet management, LTZ management, hospitals, weather, social network, etc. These data have to be accessible by an aggregator that makes queries, understanding and data integration. This is not a trivial operation since it implies the semantic understanding of the data that have to be uniformed in a single data model. A single and unified model of aggregated data allows making integrated queries to provide these data via API, and the possibility to realize services and applications. Examples of services could be those that allow geographical search, the production of suggestions based on statistical evaluations, geographic structure, similarities, etc., also on the basis of citizen behaviors on the city and with respect to the available services.
Data aggregation and provision services enable the development of apps for tourism, cultural heritage, transport and mobility, personal services, wellness, energy saving, etc. actually these opportunities are difficult to be exploited for public administrations and companies. Mainly obstacles are related to the high costs of data integration and aggregation, due the limited interoperability among data that are produced in different periods by different entities and companies.
Overview on Smart City, DISIT lab solution for beginners, 2015, Part 7: Distr...Paolo Nesi
• Smart City Concepts
• Architecture of Smart City Infrastructures
• Peripheral processors
– Data collectors and Managers
– Blog Vigilance via Natural Language Processing
– Twitter vigilance
• Data ingestion and mining
– Data Mining and smart City problematic
– Km4City: Smart City Ontology
– RDF production, reconciliation
– Parallel and distributed processing
• Reasoning and Deduction
– Smart City Engine
– Decision Support System
• Data Acting processors
– Smart City Tools and API
– Service Map and Linked Open Graph
– Mobile applications
• Projects
– SmartCity Project Sii-Mobility SCN
– SmartCity Project Coll@bora SIN
– SmartCity Project RESOLUTE H2020
– Mobile Emergency
ICARO: soluzioni e strumenti smart per avere maggiore flessibilità sul Cloud; adattare soluzioni software alle nuove esigenze cloud-based; produrre e gestire servizi a consumo: Business Process as a Service.
ICARO: Tramite modelli, strumenti e algoritmi per la gestione della configurazione e del deploy dei servizi e processi cloud; Il middleware e l’astrazione dei servizi sul cloud; l’ottimizzazione dei costi per le PMI e per la gestione del cloud.
ICARO permette: automatizzare il processo di pubblicazione e vendita delle applicazioni a consumo su cloud; automatizzare il processo di monitoraggio di basso ed alto livello e l'impostazione di strategie di smart cloud; automatizzare il controllo sulle SLA (service level agreement) in modo da associare ad evetuali disfuzioni azioni di scaling, riconfigurazione, etc.
ICARO ha sviluppato: modello descrittivo per servizi e applicazioni;
sistema automatico di configurazione;
motore di intelligence per il cloud e reasoner che prendere decisioni su configurazioni: consistenza e completezza (sulla base din un ontologia Cloud per lo Smart Cloud);
soluzione di produzione del business, config automatica;
algoritmi per il monitoraggio del comportamento di servizi e applicazioni: IaaS, PaaS, SaaS,…;
soluzione PaaS di tipo evoluto; algoritmi per la valutazione di modelli di costo e di business;
adeguamento dell’architettura su alcune applicazioni; algoritmi di ottimizzazione della gestione del cloud.
Km4City Smart City API: an integrated support for mobility servicesPaolo Nesi
AbstractThe main technical issues regarding smart city solutions are related todata gathering, aggregation, reasoning, access, and service delivering via Smart City APIs (Application Program Interfaces). Aggregated and re-conciliated data (open and private, static and real time) should be exploitable by reasoning/smart algorithms for enabling sophisticated service delivering. Different kinds of Smart City APIs enable Smart City Services and Applications, while their effectiveness depends on the architectural solutions to pass from data to services for city users and operators. To this end, a comparison of the state of the art solutions for data aggregation was performed, by putting in evidence the needs of semantic interoperable aggregated data, to provide smart services. This paper presents the work performed in the context of the Sii-Mobility national smart city project on mobility and transport integrated with services. Sii-Mobility is grounded on Km4City ontology and tools for smart city data aggregation and service production. To this end, Sii-Mobility/Km4City APIs have been compared to the state of the art solutions. Finally, the API consumption related data in the recent period are presented. Keywords smart city, smart city ontology, smart city API, smart mobility, multidomain smart city, smart services.
IEEE Smartcomp
Modelli Semantici e Gestione della Conoscenza: Social Network vs Knowledge Ma...Paolo Nesi
Modelli Semantici e Gestione della Conoscenza: Social Network vs Knowledge Management Systems. Seminario per il dottorato in Ingegneria Informatica e delle telecomunicazioni, Univ. Studi di Firenze, 2013
Sistemi Distribuiti part 5: P2P systems: from simple to distributed P2P trust...Paolo Nesi
P2P, DHT, distributed trust and certification information for DRM, P2P Health care record protection, AXMEDIS P2P, P2P for video streaming, DHT models, P2P for CDN, P2P and AXCP, monitoring P2P networks.
Smart City Strategic Forecast, SmartCity360, BratislavaPaolo Nesi
Smart City strategy, city smartening, big data amanagement,
-Taking into account results of participatory actions
-Smart city strategic plan, city agenda: prioritizing interventions
-Agreements for collaborations with main actors:
main research centers, main City Operators, etc.
-Direct collaborations on specific projects on:
ICT, Mobility, Culture, Energy, etc.
Experimenting on specific projects of the Smart City Strategic Plan
-Needs of harmonizing results and aggregating data towards dashboards
Smart Cloud Engine and Solution based on Knowledge BasePaolo Nesi
Complexity of cloud infrastructures needs models and tools for process management, configuration, scaling, elastic computing and healthiness control. This paper presents a Smart Cloud solution based on a Knowledge Base, KB, with the aim of modeling cloud resources, Service Level Agreements and their evolution, and enabling the reasoning on structures by implementing strategies of efficient smart cloud management and intelligence. The solution proposed provides formal verification tools and intelligence for cloud control. It can be easily integrated with any cloud configuration manager, cloud orchestrator, and monitoring tool, since the connections with these tools are performed by using REST calls and XML files. It has been validated in the large ICARO Cloud project with a national cloud service provider.
Km4City White Paper: Production tools for Smart City, from data to services f...Paolo Nesi
In the Smart City context, hundreds of data sets are available. Many of them are open data, accessible from local, regional, national, European public administrations, national institute of statistics, etc. These data can be static, statistical data or real time. In addition, many other data are produced by other institutions like Europeana, ECLAP, Getty, Voc, dbPedia, etc. Typically most of the data are geolocated and can be accessed as files in various formats (CSV, XLS, KMZ, JSON, XML, HTML, MySQL, ZIP, LSMA, SHP, etc.), other are accessible as Linked Data, Linked Open Data or via RDF Store end points (see dpPedia, Europeana, Senate of the Republic, Chamber of Deputies, ECLAP, km4city, etc.). Personal, private and critical data can be added to these open data. Some private data are produced by companies, like for example the position of car sharing vehicles, the position of taxis, busses, flows in the city, energy consumption data in a neighborhood, etc. Many of these data can be useful for public administrations to take decisions and to provide services. Personal data are related to a person, include personal identifications, the position of the person, its profile, etc., and need to be managed in accordance with terms of use and privacy policies. Finally, critical and personal data may be used by bad-intentioned to take actions against citizens security and infrastructures, and thus licensing and conditional access solutions are adopted.
Data are typically produced by central data producers, and many of these can provide their data in different ways and formats. Among them: traffic management systems, fleet management, LTZ management, hospitals, weather, social network, etc. These data have to be accessible by an aggregator that makes queries, understanding and data integration. This is not a trivial operation since it implies the semantic understanding of the data that have to be uniformed in a single data model. A single and unified model of aggregated data allows making integrated queries to provide these data via API, and the possibility to realize services and applications. Examples of services could be those that allow geographical search, the production of suggestions based on statistical evaluations, geographic structure, similarities, etc., also on the basis of citizen behaviors on the city and with respect to the available services.
Data aggregation and provision services enable the development of apps for tourism, cultural heritage, transport and mobility, personal services, wellness, energy saving, etc. actually these opportunities are difficult to be exploited for public administrations and companies. Mainly obstacles are related to the high costs of data integration and aggregation, due the limited interoperability among data that are produced in different periods by different entities and companies.
Overview on Smart City, DISIT lab solution for beginners, 2015, Part 7: Distr...Paolo Nesi
• Smart City Concepts
• Architecture of Smart City Infrastructures
• Peripheral processors
– Data collectors and Managers
– Blog Vigilance via Natural Language Processing
– Twitter vigilance
• Data ingestion and mining
– Data Mining and smart City problematic
– Km4City: Smart City Ontology
– RDF production, reconciliation
– Parallel and distributed processing
• Reasoning and Deduction
– Smart City Engine
– Decision Support System
• Data Acting processors
– Smart City Tools and API
– Service Map and Linked Open Graph
– Mobile applications
• Projects
– SmartCity Project Sii-Mobility SCN
– SmartCity Project Coll@bora SIN
– SmartCity Project RESOLUTE H2020
– Mobile Emergency
ICARO: soluzioni e strumenti smart per avere maggiore flessibilità sul Cloud; adattare soluzioni software alle nuove esigenze cloud-based; produrre e gestire servizi a consumo: Business Process as a Service.
ICARO: Tramite modelli, strumenti e algoritmi per la gestione della configurazione e del deploy dei servizi e processi cloud; Il middleware e l’astrazione dei servizi sul cloud; l’ottimizzazione dei costi per le PMI e per la gestione del cloud.
ICARO permette: automatizzare il processo di pubblicazione e vendita delle applicazioni a consumo su cloud; automatizzare il processo di monitoraggio di basso ed alto livello e l'impostazione di strategie di smart cloud; automatizzare il controllo sulle SLA (service level agreement) in modo da associare ad evetuali disfuzioni azioni di scaling, riconfigurazione, etc.
ICARO ha sviluppato: modello descrittivo per servizi e applicazioni;
sistema automatico di configurazione;
motore di intelligence per il cloud e reasoner che prendere decisioni su configurazioni: consistenza e completezza (sulla base din un ontologia Cloud per lo Smart Cloud);
soluzione di produzione del business, config automatica;
algoritmi per il monitoraggio del comportamento di servizi e applicazioni: IaaS, PaaS, SaaS,…;
soluzione PaaS di tipo evoluto; algoritmi per la valutazione di modelli di costo e di business;
adeguamento dell’architettura su alcune applicazioni; algoritmi di ottimizzazione della gestione del cloud.
Km4City Smart City API: an integrated support for mobility servicesPaolo Nesi
AbstractThe main technical issues regarding smart city solutions are related todata gathering, aggregation, reasoning, access, and service delivering via Smart City APIs (Application Program Interfaces). Aggregated and re-conciliated data (open and private, static and real time) should be exploitable by reasoning/smart algorithms for enabling sophisticated service delivering. Different kinds of Smart City APIs enable Smart City Services and Applications, while their effectiveness depends on the architectural solutions to pass from data to services for city users and operators. To this end, a comparison of the state of the art solutions for data aggregation was performed, by putting in evidence the needs of semantic interoperable aggregated data, to provide smart services. This paper presents the work performed in the context of the Sii-Mobility national smart city project on mobility and transport integrated with services. Sii-Mobility is grounded on Km4City ontology and tools for smart city data aggregation and service production. To this end, Sii-Mobility/Km4City APIs have been compared to the state of the art solutions. Finally, the API consumption related data in the recent period are presented. Keywords smart city, smart city ontology, smart city API, smart mobility, multidomain smart city, smart services.
IEEE Smartcomp
RESOLUTE: Resilience management guidelines and Operationalization applied to ...Paolo Nesi
--Conducting a systematic review and assessment of the state of the art of the Resilience assessment and Management concepts, national guidelines and their implementation strategies in order to develop a conceptual framework for resilience within Urban Transport Systems
--Development of European Resilience Management Guidelines (ERMG)
--Operationalize and validate the ERMG by implementing the RESOLUTE Collaborative Resilience Assessment and Management Support System (CRAMSS) for Urban Transport System (UTS) addressing Roads and Rails Infrastructures
--Enhancing resilience through improved support to human decision making processes, particularly through increased focus on the training of final users (first responders, civil protections, infrastructure managers) and population on ERMG and RESOLUTE system
--ERMG wide dissemination, acceptance and adoption at EU and Associated Countries level
DRS-7-2015 - RIA – start 1/5/2015 - end 30/4/2018 Budget 3.8M
Aggregatore di Open Data del territorio fiorentino e toscano Paolo Nesi
Dati statici e dinamici
Dati da: comune, Digital Location
Dati Real time: eventi, MIIC, Gestore (AVM, Sensori traffico, Parcheggi), LAMMA METEO, ..
Dati da UNIFI: OSIM service in LOD, RDF Store.
Dati da Social media: twitter, non ancora
Dati da Camera di Commercio: … non ancora…
Dati di flusso: da Wifi, beacom, IOT, etc.
Obiettivi e Progetti
Architettura di riferimento: in, proc, out, services
servizi, deduzioni, correlazioni, predizioni, etc.
Applicazioni: Smart City e mobilita’, energia e Smart grid, Cultural Heritage, Turismo, Decision Support Systems, Risk Assessment, Smart School, Smart Health, etc.
Progetti UNIFI: Sii-mobility SCN, Coll@bora SIN
La sfida dell’aggregazione
A che servono: Query per servizi di base e complessi:
geografiche, near to here; per comune;
NOW dati Real Time; …………………..con inferenza, per text ..
Next relevant event, …. What may happen here………
Why this event occurred….
Cloud be this feasible ???
Problematiche:
Dati di limitata interoperabilita’ semantica e qualita’ –> con molti si ottiene maggiore qualita’, l’interoperabilita’ va conquistata
Gestione grosse moli di dati, flussi, etc.
Soluzioni
Ontologia + Dati, knowledge base, inferenza, ragionamento, Ontologia Km4City
Processi di Processi di quality improvement, riconciliazione
Processi di valutazione e di supporto alle decisioni
Servizi per l’accesso ai dati
LD e RDF Store: ECLAP, OSIM, ICARO Cloud, Km4City, etc..
RDF Store, and RDF SPARQL query
LOG: Linked Open Graph, query integrate e navigazione fra store di varie istituzioni: dbpedia, Europeana, Senato, Camera, Comune, Getty, Geonames, etc. etc.
Service map
Architettura
aggregazione servizi di calcolo, parallel and distributed: Scheduler as GRID, Hadoop
NLP, quality improvement, etc.
Arricchimenti per Link e per Location
Processi di data mining, semantic computing, DSS
Km4City: Knowledge Model 4 the City: molti dati + km4city = +conoscenza e se...Paolo Nesi
Dati statici e dinamici
Obiettivi e Progetti
La sfida dell’aggregazione
Servizi per l’accesso ai dati
Lo Sviluppo di Applicazioni Web e mobile con modello Km4city
Monitoring Public Attention on Environment Issues with Twitter VigilancePaolo Nesi
DISIT Twitter Vigilance is an intelligent multi user tool for creating personal dashboards and study events and trends on Twitter that becomes a mining tool to "Twitter channels" contents. Each channel can be tuned to monitor one or more Search Queries on Twitter with a sophisticated and expressive syntax. Registered users of DISIT Twitter Vigilance tool and service are abled to:
• Create one or more channels (“canale”), as reported in the figure each channel can be tuned to monitor one or more Search Queries on Twitter with a sophisticated and expressive syntax. The simplest query can be the single keyword, tag or user.
• Create and activate multiple channels, that may use new or the same Search Queries;
• Provide public access to their channels analysis (as in the channels accessible without registration);
• Download data sets (trough API service ) for refined analysis;
• Full access at the channel history of User's content per channel, per search, per users, etc.;
• Perform visual view throughput graphs and to export them in different graphic format;
• Perform analyses at level of channel, search, users, tweets, retweets, etc.:
o trends of the Search Queries as reported in the above figure;
o distributions on population and activities of users;
o distributions about other tags/keywords;
o geographic distribution of twitters of single or multiple channels;
o distributions regarding tweet and re-tweets.
A number of research activities on these data area under development.
Functional Resonance Analysis Method based- Decision Support tool for Urban T...Paolo Nesi
Today, managing critical infrastructure resilience in smart city is a challenge that can be undertaken by adopting a new class of smart tools, which are able to integrate modeling capability with evidence driven decision support. The Resilience Decision Support tool, as presented in this article, is an innovative and powerful tool that aims at managing critical infrasctructure resilience through a more complex and expressive model based on the Functional Resonance Analysis Method and through the connection of such a model with a system thinking based decision support tool exploiting smart city data. Thanks to ResilienceDS, FRAM model becomes computable and the functional variability that is at the core of the resilience analysis can be quantified. Such quantification allows the decision support tool to compute specific strategies and recommendations for variability dampening at strategic, tactic and operational stage. The solution has been developed in the context of RESOLUTE H2020 project of the European Commission. Keywords—smart city, Functional Resonance Analysis Method; Decision Support System; Resilience; Urban Stransport System;
Km4City: A reusable example of a Metropolitan-Wide Data Platform, MAJORCITIES...Paolo Nesi
Open Source and inter-operable tools to: (i) keep city under control via personalized dashboards - monitoring services status of city operators; - monitoring and understanding the city users behaviour; - collecting moods, contributions and data from the city users; - monitoring social media for city services and events, event predictions; (ii) improve city resilience, reducing risks and decision support by: - assessing city resilience level; - improving city resilience, providing objective hints; - improving city users awareness with personal city assistants and participatory tools; (iii) transform data in value for the city: - enabling commercial and business applications; - aggregating multi-domain data and services for SMEs and city operators; - enabling integrated city services into third party web portal for all; - providing suggestion on demand services for SMEs and city operators; - accelerating and simplifying the implementation of business and service oriented Apps. Follow the Km4City City Smartener Process
Km4City, Smart City Urban Platform, From Data to Services for the Sentient Ci...Paolo Nesi
Km4City From Data to Services for the Sentient Cities
Open Source and inter-operable tools to keep city under control via personalized dashboards
- monitoring services’ status of city operators
- monitoring and understanding the city users behaviour
- collecting moods, contributions and data from the city users
- monitoring social media for city services and events, event predictions
improve city resilience, reducing risks and decision support by:
- assessing city resilience level
- improving city resilience, providing objective hints
- improving city users awareness with personal city assistants and participatory tools
transform data in value for the city:
- enabling commercial and business applications
- aggregating multi-domain data and services for SMEs and city operators
- enabling integrated city services into third party web portal for all
- providing suggestion on demand services for SMEs and city operators
- accelerating and simplifying the implementation of business and service oriented Apps
Follow the Km4City City Smartener Process
FODD 2015 Mobile App based on ServiceMap, http://www.disit.org/foddPaolo Nesi
FODD, Florence Open Data Day
Salone de’ Dugento, Palazzo Vecchio, Firenze
21/02/2015, http://www.disit.org/fodd
Ing. Ph.D Ivan Bruno
Obiettivo
Utilizzare i servizi (API REST) esposti da servicemap.disit.org
Visualizzare informazioni tempo reale / dinamiche
Realizzare un app per l’evento
Non solo una demo ma un’app estendibile e modificabile
Semplificazione
Menu configurabile
Gestione viste: una logica di gestione delle viste statiche e di quelle dinamiche da costruire a runtime sui dati JSON provenienti dalle chiamate REST via AJAX
Semplificare la gestione delle viste costruite sui dati JSON utilizzando soluzioni template-based
Rilevazione stato connessione internet del dispositivo
Notifica di anomalie (connessione assente, errori di connessioni al server....)
Portabilità su diversi dispositivi mobili
Graph Databases Lifecycle Methodology and Tool to Support Index/Store Versio...Paolo Nesi
Abstract— Graph databases are taking place in many different applications: smart city, smart cloud, smart education, etc. In most cases, the applications imply the creation of ontologies and the integration of a large set of knowledge to build a knowledge base as an RDF KB store, with ontologies, static data, historical data and real time data. Most of the RDF stores are endowed of inferential engines that materialize some knowledge as triples during indexing or querying. In these cases, deleting concepts may imply the removal and change of many triples, especially if the triples are those modeling the ontological part of the knowledge base, or are referred by many other concepts. For these solutions, the graph database versioning feature is not provided at level of the RDF stores tool, and it is quite complex and time consuming to be addressed as black box approach. In most cases the indexing is a time consuming process, and the rebuilding of the KB may imply manually edited long scripts that are error prone. Therefore, in order to solve these kinds of problems, this paper proposes a lifecycle methodology and a tool supporting versioning of indexes for RDF KB store. The solution proposed has been developed on the basis of a number of knowledge oriented projects as Sii-Mobility (smart city), RESOLUTE (smart city risk assessment), ICARO (smart cloud). Results are reported in terms of time saving and reliability.
Keywords — RDF Knowledge base versioning, graph stores versioning, RDF store management, knowledge base life cycle.
Basi di dati, SQL
Casi: Analisi della struttura di documenti
Intelligenza Artificiale
Apprendimento, clustering, alberi decisionali
Web crawling, XML, analisi dei testo
Casi di studio: OSIM
Reasoning and inferential
Big data introduction: NoSQL, graph database, ..
Social Media: user profiling, recommendations
Caso di Studio: Twitter Vigilance, sentiment analysis
Architetture parallele
Casi di studio: smart city
Casi di studio: Smart Cloud
Smart City and Open Data Projects and tools of DISIT LabPaolo Nesi
Current research topics
• Social media, collaborative work, Mobile computing, OpenData, LOD
• SmartCity, BigData, data analytics
• Railway signaling, autonomous driving systems, formal methods
• Cloud Computing, grid computing, smart cloud
• Data Mining, Knowledge Acceleration, natural language processing
Main research results
• Knowledge Management and Natural Language Processing: OSIM, CoSkoSAM
• Content and Protection Management, grid computing: AXMEDIS AXCP
• Social Media, recommendations and tool: ECLAP.eu, MyStoryPlayer, Social Graph, IPR Wizard…
• Mobile Computing: Mobile Medicine, Mobile Emergency, etc.….
• Music Transcode, winner of MIREX for piano
• Awards: IEEE ICECCS, DMS, Italia degli Innovatori, etc.
Main sources of funding
• European Commission: ECLAP (social media, Cultural Heritage, open data), AXMEDIS (DRM, protection, automation e grid computing), WEDELMUSIC, IMAESTRO, VARIAZIONI, IMUTUS, MUSICNETWORK, MOODS, MUPAAC, OFCOMP, etc. ……
• Italian Ministry: Smart Cities COLL@BORA (collaborative work, social media), FIRB e PRIN
• Regional: SACVAR (knowledge mining and reasoning), TRACE‐IT (Railway signalling), RAISSS (Railway signalling), ICARO (cloud)
• Fondations: MatchMaking (NLP), OSIM (Knowledge Acceleration, NLP)
DISIT Potential challenges and interests
DISIT is interested in participating in the next calls of the European Commission and in particular for:
• Working on open data and linked open data for smart city, smart cloud, smart manufacturing, smart museum, etc.
• Creating semantic models and reasoning engines
• Creating data mining and natural language processing tools as SACVAR/OSIM
• Working on defining big data solutions and infrastructures
• Working on data analytics algorithms computing:
• Predictions and trends,
• unexpected correlations,
• data inconsistencies and incompleteness,
• etc.
Twitter Vigilance: Modelli e Strumenti per l’Analisi e lo Studio di Dati Soci...Paolo Nesi
Twitter Vigilance: le analisi http://www.disit.org
Analisi e caratterizzazione della comunicazione
Percezione sociale, eventi pubblici, naturali..
Scoprire, identificare e calcolare
Nascita / crescita di nuove occorrenze in tempo reale: eventi, fatti, meteo, condizioni critiche, etc.
Supporto alla decisioni, ridurre i tempi di reazione, valutare la percezione, ridurre i costi, incrementare la resilienza come capacità di reagire, diagnosi precoce
Chi influenza la comunicazione, le comunità e come: i pusher, gli attori, i follower, le sorgenti, etc.
Predizione su eventi periodici, per esempio presenze ad eventi, presenze sui canali televisivi, vendite aziende, etc.
Misure indirette basate sulla popolazione: rischio sicurezza, degrado, neve, grandine, vento, fallimenti, etc.
Km4City: una soluzione aperta per erogare servizi Smart CityPaolo Nesi
Km4City: Integrated Urban Platform, Open Source
Aggregate & integrate data
Multiple protocols from urban operators, ....
open data, IOT, sensors, internet of everything, cloud, mobile devices, Wi-Fi, social media, ...
Data Exploitation performing
predictions, reasoning, business intelligence, ..
users behavior analysis, decision support system, ..
Control Room, Real Time Monitoring tools, ….
Produce value from data enabling to
Stimulate virtuous behavior, influence City Users!
Put in action CITY Strategies
Il “Grillo parlante” è un’applicazione java che consente la ricezione e l’invio di notifiche, messaggi, file di testo e file multimediali. Inoltre è possibile creare e gestire gruppi informali, gestire appuntamenti e, all'interno dell' edificio universitario U14 - Bicocca, è possibile localizzazione gli utenti cercati.
Social Media - Introduzione al Corso [a.a. 2014-2015] - UniToAgnese Vellar
Introduzione al corso per gli studenti delle Lauree Magistrali di Comunicazione Pubblica e Politica e Comunicazione ICT e Media - Università degli Studi di Torino http://goo.gl/B6vE6M
Intervento di Maurizio Galliano, DYRECTA - Living Lab ICT e E-LEARNING2.0
OPEN DAY - COMPETENZE DIGITALI
Sala Convegni Pad. 152 Regione Puglia Fiera del levante Bari
15 maggio 2015 ore 9.30
RESOLUTE: Resilience management guidelines and Operationalization applied to ...Paolo Nesi
--Conducting a systematic review and assessment of the state of the art of the Resilience assessment and Management concepts, national guidelines and their implementation strategies in order to develop a conceptual framework for resilience within Urban Transport Systems
--Development of European Resilience Management Guidelines (ERMG)
--Operationalize and validate the ERMG by implementing the RESOLUTE Collaborative Resilience Assessment and Management Support System (CRAMSS) for Urban Transport System (UTS) addressing Roads and Rails Infrastructures
--Enhancing resilience through improved support to human decision making processes, particularly through increased focus on the training of final users (first responders, civil protections, infrastructure managers) and population on ERMG and RESOLUTE system
--ERMG wide dissemination, acceptance and adoption at EU and Associated Countries level
DRS-7-2015 - RIA – start 1/5/2015 - end 30/4/2018 Budget 3.8M
Aggregatore di Open Data del territorio fiorentino e toscano Paolo Nesi
Dati statici e dinamici
Dati da: comune, Digital Location
Dati Real time: eventi, MIIC, Gestore (AVM, Sensori traffico, Parcheggi), LAMMA METEO, ..
Dati da UNIFI: OSIM service in LOD, RDF Store.
Dati da Social media: twitter, non ancora
Dati da Camera di Commercio: … non ancora…
Dati di flusso: da Wifi, beacom, IOT, etc.
Obiettivi e Progetti
Architettura di riferimento: in, proc, out, services
servizi, deduzioni, correlazioni, predizioni, etc.
Applicazioni: Smart City e mobilita’, energia e Smart grid, Cultural Heritage, Turismo, Decision Support Systems, Risk Assessment, Smart School, Smart Health, etc.
Progetti UNIFI: Sii-mobility SCN, Coll@bora SIN
La sfida dell’aggregazione
A che servono: Query per servizi di base e complessi:
geografiche, near to here; per comune;
NOW dati Real Time; …………………..con inferenza, per text ..
Next relevant event, …. What may happen here………
Why this event occurred….
Cloud be this feasible ???
Problematiche:
Dati di limitata interoperabilita’ semantica e qualita’ –> con molti si ottiene maggiore qualita’, l’interoperabilita’ va conquistata
Gestione grosse moli di dati, flussi, etc.
Soluzioni
Ontologia + Dati, knowledge base, inferenza, ragionamento, Ontologia Km4City
Processi di Processi di quality improvement, riconciliazione
Processi di valutazione e di supporto alle decisioni
Servizi per l’accesso ai dati
LD e RDF Store: ECLAP, OSIM, ICARO Cloud, Km4City, etc..
RDF Store, and RDF SPARQL query
LOG: Linked Open Graph, query integrate e navigazione fra store di varie istituzioni: dbpedia, Europeana, Senato, Camera, Comune, Getty, Geonames, etc. etc.
Service map
Architettura
aggregazione servizi di calcolo, parallel and distributed: Scheduler as GRID, Hadoop
NLP, quality improvement, etc.
Arricchimenti per Link e per Location
Processi di data mining, semantic computing, DSS
Km4City: Knowledge Model 4 the City: molti dati + km4city = +conoscenza e se...Paolo Nesi
Dati statici e dinamici
Obiettivi e Progetti
La sfida dell’aggregazione
Servizi per l’accesso ai dati
Lo Sviluppo di Applicazioni Web e mobile con modello Km4city
Monitoring Public Attention on Environment Issues with Twitter VigilancePaolo Nesi
DISIT Twitter Vigilance is an intelligent multi user tool for creating personal dashboards and study events and trends on Twitter that becomes a mining tool to "Twitter channels" contents. Each channel can be tuned to monitor one or more Search Queries on Twitter with a sophisticated and expressive syntax. Registered users of DISIT Twitter Vigilance tool and service are abled to:
• Create one or more channels (“canale”), as reported in the figure each channel can be tuned to monitor one or more Search Queries on Twitter with a sophisticated and expressive syntax. The simplest query can be the single keyword, tag or user.
• Create and activate multiple channels, that may use new or the same Search Queries;
• Provide public access to their channels analysis (as in the channels accessible without registration);
• Download data sets (trough API service ) for refined analysis;
• Full access at the channel history of User's content per channel, per search, per users, etc.;
• Perform visual view throughput graphs and to export them in different graphic format;
• Perform analyses at level of channel, search, users, tweets, retweets, etc.:
o trends of the Search Queries as reported in the above figure;
o distributions on population and activities of users;
o distributions about other tags/keywords;
o geographic distribution of twitters of single or multiple channels;
o distributions regarding tweet and re-tweets.
A number of research activities on these data area under development.
Functional Resonance Analysis Method based- Decision Support tool for Urban T...Paolo Nesi
Today, managing critical infrastructure resilience in smart city is a challenge that can be undertaken by adopting a new class of smart tools, which are able to integrate modeling capability with evidence driven decision support. The Resilience Decision Support tool, as presented in this article, is an innovative and powerful tool that aims at managing critical infrasctructure resilience through a more complex and expressive model based on the Functional Resonance Analysis Method and through the connection of such a model with a system thinking based decision support tool exploiting smart city data. Thanks to ResilienceDS, FRAM model becomes computable and the functional variability that is at the core of the resilience analysis can be quantified. Such quantification allows the decision support tool to compute specific strategies and recommendations for variability dampening at strategic, tactic and operational stage. The solution has been developed in the context of RESOLUTE H2020 project of the European Commission. Keywords—smart city, Functional Resonance Analysis Method; Decision Support System; Resilience; Urban Stransport System;
Km4City: A reusable example of a Metropolitan-Wide Data Platform, MAJORCITIES...Paolo Nesi
Open Source and inter-operable tools to: (i) keep city under control via personalized dashboards - monitoring services status of city operators; - monitoring and understanding the city users behaviour; - collecting moods, contributions and data from the city users; - monitoring social media for city services and events, event predictions; (ii) improve city resilience, reducing risks and decision support by: - assessing city resilience level; - improving city resilience, providing objective hints; - improving city users awareness with personal city assistants and participatory tools; (iii) transform data in value for the city: - enabling commercial and business applications; - aggregating multi-domain data and services for SMEs and city operators; - enabling integrated city services into third party web portal for all; - providing suggestion on demand services for SMEs and city operators; - accelerating and simplifying the implementation of business and service oriented Apps. Follow the Km4City City Smartener Process
Km4City, Smart City Urban Platform, From Data to Services for the Sentient Ci...Paolo Nesi
Km4City From Data to Services for the Sentient Cities
Open Source and inter-operable tools to keep city under control via personalized dashboards
- monitoring services’ status of city operators
- monitoring and understanding the city users behaviour
- collecting moods, contributions and data from the city users
- monitoring social media for city services and events, event predictions
improve city resilience, reducing risks and decision support by:
- assessing city resilience level
- improving city resilience, providing objective hints
- improving city users awareness with personal city assistants and participatory tools
transform data in value for the city:
- enabling commercial and business applications
- aggregating multi-domain data and services for SMEs and city operators
- enabling integrated city services into third party web portal for all
- providing suggestion on demand services for SMEs and city operators
- accelerating and simplifying the implementation of business and service oriented Apps
Follow the Km4City City Smartener Process
FODD 2015 Mobile App based on ServiceMap, http://www.disit.org/foddPaolo Nesi
FODD, Florence Open Data Day
Salone de’ Dugento, Palazzo Vecchio, Firenze
21/02/2015, http://www.disit.org/fodd
Ing. Ph.D Ivan Bruno
Obiettivo
Utilizzare i servizi (API REST) esposti da servicemap.disit.org
Visualizzare informazioni tempo reale / dinamiche
Realizzare un app per l’evento
Non solo una demo ma un’app estendibile e modificabile
Semplificazione
Menu configurabile
Gestione viste: una logica di gestione delle viste statiche e di quelle dinamiche da costruire a runtime sui dati JSON provenienti dalle chiamate REST via AJAX
Semplificare la gestione delle viste costruite sui dati JSON utilizzando soluzioni template-based
Rilevazione stato connessione internet del dispositivo
Notifica di anomalie (connessione assente, errori di connessioni al server....)
Portabilità su diversi dispositivi mobili
Graph Databases Lifecycle Methodology and Tool to Support Index/Store Versio...Paolo Nesi
Abstract— Graph databases are taking place in many different applications: smart city, smart cloud, smart education, etc. In most cases, the applications imply the creation of ontologies and the integration of a large set of knowledge to build a knowledge base as an RDF KB store, with ontologies, static data, historical data and real time data. Most of the RDF stores are endowed of inferential engines that materialize some knowledge as triples during indexing or querying. In these cases, deleting concepts may imply the removal and change of many triples, especially if the triples are those modeling the ontological part of the knowledge base, or are referred by many other concepts. For these solutions, the graph database versioning feature is not provided at level of the RDF stores tool, and it is quite complex and time consuming to be addressed as black box approach. In most cases the indexing is a time consuming process, and the rebuilding of the KB may imply manually edited long scripts that are error prone. Therefore, in order to solve these kinds of problems, this paper proposes a lifecycle methodology and a tool supporting versioning of indexes for RDF KB store. The solution proposed has been developed on the basis of a number of knowledge oriented projects as Sii-Mobility (smart city), RESOLUTE (smart city risk assessment), ICARO (smart cloud). Results are reported in terms of time saving and reliability.
Keywords — RDF Knowledge base versioning, graph stores versioning, RDF store management, knowledge base life cycle.
Basi di dati, SQL
Casi: Analisi della struttura di documenti
Intelligenza Artificiale
Apprendimento, clustering, alberi decisionali
Web crawling, XML, analisi dei testo
Casi di studio: OSIM
Reasoning and inferential
Big data introduction: NoSQL, graph database, ..
Social Media: user profiling, recommendations
Caso di Studio: Twitter Vigilance, sentiment analysis
Architetture parallele
Casi di studio: smart city
Casi di studio: Smart Cloud
Smart City and Open Data Projects and tools of DISIT LabPaolo Nesi
Current research topics
• Social media, collaborative work, Mobile computing, OpenData, LOD
• SmartCity, BigData, data analytics
• Railway signaling, autonomous driving systems, formal methods
• Cloud Computing, grid computing, smart cloud
• Data Mining, Knowledge Acceleration, natural language processing
Main research results
• Knowledge Management and Natural Language Processing: OSIM, CoSkoSAM
• Content and Protection Management, grid computing: AXMEDIS AXCP
• Social Media, recommendations and tool: ECLAP.eu, MyStoryPlayer, Social Graph, IPR Wizard…
• Mobile Computing: Mobile Medicine, Mobile Emergency, etc.….
• Music Transcode, winner of MIREX for piano
• Awards: IEEE ICECCS, DMS, Italia degli Innovatori, etc.
Main sources of funding
• European Commission: ECLAP (social media, Cultural Heritage, open data), AXMEDIS (DRM, protection, automation e grid computing), WEDELMUSIC, IMAESTRO, VARIAZIONI, IMUTUS, MUSICNETWORK, MOODS, MUPAAC, OFCOMP, etc. ……
• Italian Ministry: Smart Cities COLL@BORA (collaborative work, social media), FIRB e PRIN
• Regional: SACVAR (knowledge mining and reasoning), TRACE‐IT (Railway signalling), RAISSS (Railway signalling), ICARO (cloud)
• Fondations: MatchMaking (NLP), OSIM (Knowledge Acceleration, NLP)
DISIT Potential challenges and interests
DISIT is interested in participating in the next calls of the European Commission and in particular for:
• Working on open data and linked open data for smart city, smart cloud, smart manufacturing, smart museum, etc.
• Creating semantic models and reasoning engines
• Creating data mining and natural language processing tools as SACVAR/OSIM
• Working on defining big data solutions and infrastructures
• Working on data analytics algorithms computing:
• Predictions and trends,
• unexpected correlations,
• data inconsistencies and incompleteness,
• etc.
Twitter Vigilance: Modelli e Strumenti per l’Analisi e lo Studio di Dati Soci...Paolo Nesi
Twitter Vigilance: le analisi http://www.disit.org
Analisi e caratterizzazione della comunicazione
Percezione sociale, eventi pubblici, naturali..
Scoprire, identificare e calcolare
Nascita / crescita di nuove occorrenze in tempo reale: eventi, fatti, meteo, condizioni critiche, etc.
Supporto alla decisioni, ridurre i tempi di reazione, valutare la percezione, ridurre i costi, incrementare la resilienza come capacità di reagire, diagnosi precoce
Chi influenza la comunicazione, le comunità e come: i pusher, gli attori, i follower, le sorgenti, etc.
Predizione su eventi periodici, per esempio presenze ad eventi, presenze sui canali televisivi, vendite aziende, etc.
Misure indirette basate sulla popolazione: rischio sicurezza, degrado, neve, grandine, vento, fallimenti, etc.
Km4City: una soluzione aperta per erogare servizi Smart CityPaolo Nesi
Km4City: Integrated Urban Platform, Open Source
Aggregate & integrate data
Multiple protocols from urban operators, ....
open data, IOT, sensors, internet of everything, cloud, mobile devices, Wi-Fi, social media, ...
Data Exploitation performing
predictions, reasoning, business intelligence, ..
users behavior analysis, decision support system, ..
Control Room, Real Time Monitoring tools, ….
Produce value from data enabling to
Stimulate virtuous behavior, influence City Users!
Put in action CITY Strategies
Il “Grillo parlante” è un’applicazione java che consente la ricezione e l’invio di notifiche, messaggi, file di testo e file multimediali. Inoltre è possibile creare e gestire gruppi informali, gestire appuntamenti e, all'interno dell' edificio universitario U14 - Bicocca, è possibile localizzazione gli utenti cercati.
Social Media - Introduzione al Corso [a.a. 2014-2015] - UniToAgnese Vellar
Introduzione al corso per gli studenti delle Lauree Magistrali di Comunicazione Pubblica e Politica e Comunicazione ICT e Media - Università degli Studi di Torino http://goo.gl/B6vE6M
Intervento di Maurizio Galliano, DYRECTA - Living Lab ICT e E-LEARNING2.0
OPEN DAY - COMPETENZE DIGITALI
Sala Convegni Pad. 152 Regione Puglia Fiera del levante Bari
15 maggio 2015 ore 9.30
Per apprendimento online (noto anche come teleapprendimento, o con il termine inglese E-learning) s'intende l'uso delle tecnologie multimediali e di Internet per migliorare la qualità dell'apprendimento facilitando l'accesso alle risorse e ai servizi, così come anche agli scambi in remoto e alla collaborazione a distanza
OpenCoesione come strumento per favorire la partecipazioneOpenCoesione
Presentazione di Luigi Reggi (DPS) durante il Seminario di formazione SSAI "Conoscere le politiche di coesione per lo sviluppo dei territori" (Roma, 1-5 luglio 2013)
Presentazione CSI Piemonte - Fossano 11 dicembre 2014 - parte 2Giuly Bonello
Aspetti territoriali dell'uso di dati informativi nelle PA: esperienze e opportunità - presentazione di CSI Piemonte (G. Bonello e M. Cavagnoli) al Comune di Fossano (11 dicembre 2014)
Quali sono i nuovi trends del mobile learning? Come sviluppare competenze digitali in tutta l'organizzazione attraverso un’app? Quali spunti per la formazione ci possono dare le app per apprendere le lingue e fare sport?
I trends del mobile learning e i principi didattici e di engagement sono alla base delle app skilla; scoprilo grazie alla presentazione dell'app DigitalJourney e alla dimostrazione del suo funzionamento. Importante l’approfondimento sul ruolo delle app nell'ecosistema digitale di apprendimento dell'organizzazione.
Esistono nuovi modelli e nuovi strumenti per guidare i cambiamenti culturali e tecnologici all'interno dell'organizzazione; quali sono?
CONTENUTI:
- Il mobile learning in numeri
- Vantaggi e barriere del mobile learning
- DigitalJourney. L’app per guidare le persone nella trasformazione digitale
Collegandoti al seguente link, potrai visionare l’abstract video del seminario online:
https://youtu.be/a25-3tXOIOs
Il Piano di comunicazione del progetto Partecipa.net
Estrazione e Deduzione della Conoscenza via Modelli Semantici: From Social Network to Smart City
1. Estrazione e Deduzione della
Conoscenza via Modelli Semantici:
From Social Network to Smart City
seminario per il Corso di Dottorato 2014
Prof. Paolo Nesi
Dipartimento di Ingegneria dell’Informazione
University of Florence
Via S. Marta 3, 50139, Firenze, Italy
tel: +39-055-4796523, fax: +39-055-4796363
DISIT Lab
Paolo.nesi@unifi.it
http://www.disit.dinfo.unifi.it/
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
1
2. I Department
of Information Engineering (DINFO)
Distributed Data Intelligence and Technologies
I Distributed Systems and Internet Technologies
I http://www.disit.dinfo.unifi.it
I DISIT,
I http://www.disit.dinfo.unifi.it
I Knowledge
Acceleration
I Data
Analytics, Big data
I Social Media
I Mobile
I Cloud
Computing
I Smart
Cities
and Grid Computing
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
2
3. Department of Information Engineering (DINFO)
DISIT, Distributed Data Intelligence and Technologies
Distributed Systems and Internet Technologies
http://www.disit.dinfo.unifi.it
Staff
http://www.disit.dinfo.unifi.it
• Prof. Paolo Nesi, paolo.nesi@unifi.it
• More than 20 among postdocs, PhD students and fellows
Current research topics
• Social media, collaborative work, Mobile computing, OpenData, LOD
• SmartCity, BigData, data analytics
• Railway signaling, autonomous driving systems, formal methods
• Cloud Computing, grid computing, smart cloud
• Data Mining, Knowledge Acceleration, natural language processing
Main research results
• Knowledge Management and Natural Language Processing: OSIM, CoSkoSAM
• Content and Protection Management, grid computing: AXMEDIS AXCP
• Social Media, recommendations and tool: ECLAP.eu, MyStoryPlayer, Social Graph, IPR Wizard…
• Mobile Computing: Mobile Medicine, Mobile Emergency, etc.….
• Music Transcode, winner of MIREX for piano
• Awards: IEEE ICECCS, DMS, Italia degli Innovatori, etc.
Main sources of funding
• European Commission: ECLAP (social media, CH), AXMEDIS (DRM, protection, automation e grid
computing), WEDELMUSIC, IMAESTRO, VARIAZIONI, IMUTUS, MUSICNETWORK, MOODS,
MUPAAC, OFCOMP, etc. ……
• Italian Ministry: Smart Cities COLL@BORA (collaborative work, social media), FIRB e PRIN
• Regional: SACVAR (knowledge mining and reasoning), TRACE‐IT (Railway signalling), RAISSS
(Railway signalling), ICARO (cloud)
• Fondations: MatchMaking (NLP), OSIM (Knowledge Acceleration, NLP)
I Dottorato,
Univ. Firenze, Paolo Nesi 2013-2014
I3
4. Struttura del Seminario
G
G
G
G
G
G
G
G
Social Networking and knowledge
Semantic and Social Networks
Recommendations and Suggestions
Natural Language Processing System
Knowledge Representation System
Reasoning System
Sistema OSIM
Smart Cities
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
4
5. Profilo degli utenti
Informazioni
statiche:
Informazioni dinamiche:
•Informazioni generali:
•Lista di oggetti preferiti
•nome, cognome, sesso,
•Lista di amici
• foto, data di nascita,
•Lista gruppi
•descrizione personale,
•Voti positivi ad oggetti
•località di provenienza (ISO 3166),
•Commenti ad oggetti
•Nazione
•…
•Suddivisione
•Provincia
•…
•lingue parlate (ISO 369)
•Informazioni sulle preferenze
•Informazioni di contatto:
sulla base delle visualizzazioni
•lista di contatti di instant messaging
degli oggetti
•Scuola e Lavoro:
•Format
•scelta del livello scolastico,
•Type
•nome della scuola,
•Taxonomy
•tipo di lavoro,
•nome del posto di lavoro
•Interessi:
•Vettore contenente la lista di valori del campo Type degli oggetti
scelti dall’utente
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
5
6. Content Searching in Social Network
G
Traditional Classification based on
Metadata
Faceted search
G
G
G
G
G
Taxonomies
Free Tags, as Folksonomy
Geotagging, GPS data
Annotations
Votes
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
6
7. User Profile Problems
G
Different data types:
Numbers: age, votes, #kids, ..
Enumerates/symbolic: language, nationality, etc.
G
Multiple Values / Selections:
languages, nationalities, preferences, etc…
G
Non-Symmetrical Distances, for instance:
Preferences: Dim ({Pref(A)}) ≠ Dim ({Pref(B)})
G
Dynamic information
related computational complexity
G
Different Languages of comments, descriptions,
Language processing and understanding
Dictionaries, Semantics, Taxonomy, etc.
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
7
8. friendship propagation
G
G
User links and friendship propagation….
Mechanisms for invitation
User A invites N Users
Among these N Users, M Accept the invitation
G
Viral Indicator
If M > N a mechanism of viral grow is started
It can exponentially grow up or to simply produce a small
pike
G
Users have:
Direct Friends----------------- for example: 90
Indirect Friend of different levels ---------: level 1: 900
Friends via groups (see LinkedIn) ---------: 14000
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
8
9. Struttura del Seminario
G
G
Social Networking and knowledge
Semantic and Social Networks
G
G
G
G
G
G
G
Semantic Modeling
Recommendations and Suggestions
Natural Language Processing System
Knowledge Representation System
Reasoning System
Sistema OSIM
Smart Cities
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
9
10. Semantic Descriptors and info 1/3
G
user profile descriptions collected via user registration and
dynamically on the basis of user actions, migrated also on the
mobile:
selected content, performed queries,
preferred content, suggested content, etc.;
G
G
relationships among users/colleagues (similarly to
friendships, group joining) that impact on the user profile and
are created via registration, by inviting colleagues, etc.;
user groups descriptors and their related discussion forums
and web pages (with taxonomic descriptors and text);
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
10
11. Semantic Descriptors and info 2/3
G
G
G
content descriptors for simple and complex content, web
pages, forums, comments, etc.;
device capabilities for formal description of any acceptable
content format and parameters, CPU capabilities, memory
space, SSD space;
votes and comments on contents, forums, web pages,
etc., which are dynamic information related to users;
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
11
12. Semantic Descriptors and info 3/3
G
G
G
lists of elements marked as preferred by users, which are
dynamic information related to users;
downloads and play/executions of simple and/or complex
content on PC and mobiles, to keep trace of user actions as
references to played content, which are dynamic information
related to users preferences;
uploads and publishing of user provided content on the
portal (only for registered users, and supervised by the
administrator of the group). Each Content element has its
own static metadata, descriptors and taxonomy; while the
related action of upload is a dynamic information associated
with the User who performed it. In addition, Content elements
can be associated with Groups.
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
12
13. Group Descriptors
G
Groups of users they may have specific
descriptors and those inherited by the users:
G
static aspects of the groups such as:
objectives, topics, web pages, keywords, taxonomy, etc.;
G
dynamic aspects related to:
users belonging to the group; users may: join and leave the
group, be more or less active over time;
content associated with the group: files, comments, etc.,
with their taxonomical classification, metadata and
descriptors.
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
13
14. Content Descriptors
G
Static aspects : more relevant since the content description is
typically not changing over time. They are:
metadata, keywords extracted from description, comments, etc.;
technical description (as the Format in the following): audio, video,
document, cross media, image,..;
content semantic descriptors such as: rhythm, color, etc.; genre,
called Type in the following;
groups to which the content has been associated with;
taxonomies classification to which the content has been
associated, taking into account also the general taxonomy;
G
dynamic aspects are marginally changed and may be related
to:
user’s votes, user’s comments;
number of votes, comments, download, direct recommendations,
etc.;
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
14
15. Semantic Descriptors
G
Modeling descriptors with formalisms:
XML
MPEG-7, metamodel for descriptors and descriptors
MPEG-21: item descriptor and/or package
G
Audio, Video, images:
Low level fingerprint/descriptors
Hash, MD5, etc.
High level fingerprint/descriptors
Genre, rhythms, color, scenes/movements, etc.
Evolution of them along the time, along the file
G
Documents:
Keywords extractions, multilingual agnostic, …
Summarization
Paragraphs modeling and descriptions
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
15
16. Usage/Prod of Semantic Information
G
G
G
G
G
G
G
G
content ingestion. semantic tagging while technical descriptors about
digital resources are added during the automated adaptation and icon
production;
repurposing and publication for several kinds of end-user
devices
extraction of semantic technical descriptors from simple and
complex essences,
content indexing to prepare and accelerate the process of search.
packaging content and semantics into MPEG-21/AXMEDIS binary
format: integrating digital essences with metadata and descriptors
exporting content to other databases, or posting them on other
social networks or portals, publishing on P2P networks
estimating similarities among users, objects/content, to pose the
basis of generating suggestions and reasoning;
producing suggestions about potential colleagues, interesting
content, and groups;
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
16
17. Content Enrichment
G
G
The content, UGC, reaches the Social Network with partial
information
Content Enrichment is needed to get enough semantic
information for
indexing/querying and producing suggestions
G
Content enrichment may be performed by:
Addition/Extraction of semantic descriptors
Multilingual translation for metadata
Addition of annotations, textual and audiovisual
Association of SKOS/taxonomical terms
Association of Tags folksonomy
Comments, rating, citations, etc.
Creation of Aggregations: collection, courses, play lists
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
17
18. Extraction of semantic
descriptors
G
Technical Information
duration, resolution, size, dimension, video rate, sample rate and
size, file format, MIME type, number of included files, file
extension, etc.
libraries or tools can be used to extract information: FFMPEG for
video and audio, ImageMagik for images, etc.
G
Context information:
summary and extract keywords;
Video processing to: segment in major scenes, understand
them, identify objects, colors, etc.
audio processing to extract tonality, rhythm, etc.
Images processing to extract contained objects, etc.
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
18
19. High Level Reasoning Semantic Computing, 1/2
G
Linguistic processing: assessment of intentions, understanding
Extraction of positive/negative impressions
Technical instruments:
Ontology production, integration, augmentation
Ontology merging, engines
Processing OWL
Triple database, Semantic SQL
G
Semantic meaning of high level information
Dictionaries: to compare/infer multilingual keywords
Folksonomies: production of free keywords
Taxonomies: specialization relationships
Ontology: a range of relationships
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
19
20. High Level Reasoning Semantic Computing, 2/2
G
Taking decision on the basis of Descriptors and their relationships
Technical instruments:
Taking decision engines
inferential engines such as Jena,
rules based systems,
script-based rules,
constraint programming,
First logic, temporal logic engine, etc.
G
Recommendations/suggestions, production of
Technical instruments:
Clustering among elements: content, users, groups, ..
• on the basis of distances/similarities among descriptors
Clustering models: K-means, k-medoid, hierarchical clustering
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
20
22. Alternative, traduzioni
Horse, equino, …..
isa
Animale
Bretone
Cavallo
isa
Normanno
lupo
hardware
Computer
Science
sys..
isa
Sistema
Software
Related
task
Referenced
economia
Modello
economico
collabora
tive
Process
Assisted
Competitive
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
22
23. FOAF: Friend of a Friend
G
Friend of a Friend (FOAF):
A format for supporting description of people and
their relationships
a vocabulary in OWL for sharing personal and
social network information on the Semantic Web
Based on AAA principle:
Anyone can say anything about any topic
G
Modeling Information:
Organization at which people belong
Documents that people have created/co-authored
Images that depict people
Interests/skill of people,…
…..
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
23
25. Ontologie
G
G
G
L'ontologia è una specificazione formale esplicita
di una concettualizzazione di un dominio
Rappresentano:
Concetti e oggetti: modelli, categorie,
proprietà,..
Relazioni fra concetti e fra relazioni
Idealmente mirano a modellare in modo
“esaustivo” un dominio
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
25
27. Base Ontologica
G
G
Si può formalizzare in OWL (ontology
web
language), XML (Extensible Markup
Language)
Unifica/Generalizza modelli come:
Tassonomie
Tesauri
Vocabolari
SKOS: simple knowledge organization system
FOAF: friend of a friend
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
27
28. Base Ontologica
G
G
Le ontologie sono specifiche di un dominio
Spesso prodotte in team e formalizzate in OWL, vi
sono strumenti di:
Editing, e.g., Protégé
Database semantici, e.g., Sesame in RDF
(Resource Description Language)
Inferenza su database
Query semantiche, per esempio formalizzati in
SPARQL (Simple Protocol and RDF Query
Language)
….
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
28
29. Base di Conoscenza
G
G
Si può vedere come: Ontologia + istanze
Le istanze dei concetti, delle relazioni,
popolano la base di conoscenza
connettendosi all’ontologia con la relazione
di instance-of (io), e.g.:
il documento afkagf.pdf connesso al nodo
Analgesia della tassonomia MobMed
Carlo Rossi e’ un Paziente
Carlo Rossi e’ figlio di Giovanni Rossi
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
29
30. Struttura del Seminario
G
G
G
Social Networking and knowledge
Semantic and Social Networks
Recommendations and Suggestions
G
G
G
G
G
G
G
G
G
G
G
Raccomandazioni / suggerimenti
Metrics Similarity Distances
Clustering algorithms comparison
Performances, Incremental Clustering
Suggerimenti UU an improvement
Validazione del modello di suggerimento
Natural Language Processing System
Knowledge Representation System
Reasoning System
Sistema OSIM
Smart Cities
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
30
31. Recommendations
G
Different Recommendations/Suggestions
G
U U: a user to another user on the basis of user profile
O U: an object to a user on the basis of user profile
O O: an object on the basis of a played object of a user
G U: a group to a user on the basis of user profile
Etc…
Objects can be:
Advertising, Content, Events, etc.
Some of them may have specific descriptors…
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
31
32. Different Recommendations
G
FOR YOU: Suggesting Users to another Users since they
have similar preferences
like/prefer what you like/prefer
are friends of your friends
are in one or more of the your groups
are new of the Social Network
are the most linked, grouped, active
etc.
G
FOR THE SN: Suggesting Users to another Users since they
are important for the SN and do not have to left alone, the new entry
are the only contact path for Connecting a remote group, if the path is
left a peripheral group will be completely disjoined with respect to the
rest of the SN
…
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
32
33. Different Recommendations
G
FOR YOU: Suggested objects/contents/events/groups since they
are the less, most viewed, most played, most played in your group, ..
are similar to your highest voted/ranked objects
are similar to what you usually play, pay, print, upload, etc.
The most played/../voted in absolute
The most played/../voted in the last Month/Day, week, etc…
The most played/../voted in your area, country, group, etc..
are new for the SN
belongs to the preferred of your friends, …
have been posted/commented by your friends, in your group, …
have been recommended by a your friend
G
FOR BUSINESS: Suggested objects/…./groups since they
are new for the SN, and thus are new for the market/business of the SN
are commercially proposed and have to be commercially promoted for the
business of the SN
belong on the log tail of the content distribution/usage
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
33
34. Recommendations
User
Suggested elements
Users
Proposing to a user
possible colleagues /
friends
Recipient of the suggestions
Content
(played by a user)
--no sense--
Contents
Proposing to a user
possible interesting
contents
Proposing at a play of a
content similar content
items
Groups
Proposing to a user
possible interesting
groups
Proposing at a play of a
content possible interesting
groups in which similar
contents are discussed
Ads
Proposing to a user
Proposing at a play of a
possible interesting ads content the possible
interesting ads
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
Group
(leader or members)
Proposing at a group
responsible possible
interested colleagues to be
invited
Proposing at a group
members possible interesting
content
(not much different with
respect to C-C combination)
--no sense--
Proposing at a/all group
member/s possible
interesting ads
34
35. Similarity Distance
G
G
G
G
The simplest solution for the recommendations/suggestion is
to estimate the closest Users or Objects with respect to the
reference User/Object
The estimation of the closest entity between two entities
described with multiple symbolic description is an instance of
multidomain symbolic similarity distance among their
descriptors.
We can suppose for a while to have the possibility of
estimating the similarity distance among descriptors.
Some indexing tools, such as Lucene/Solr, may help in doing
this with a query based on information of the reference
user/object.
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
35
36. Complexity of Recommendation 1/2
G
G
G
Each day:
N new users reach the SN
The SN has 1 Million of users: U=10^6
The SN has to suggest the possible friends to the new N
users immediately:
Complexity is an O(NU)
N*U distances should be estimated in real time/per day
If N=10^6 such as on YouTube
Thus: 10^12 estimations of 10ms,
10^10s which are 317 years !!!
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
36
37. Complexity of Recommendation 2/2
G
G
G
Each day:
M new UGC items are uploaded on the SN,
The SN has
1 Million of content: C=10^6
1 Million of users: U=10^6
The SN has to estimate the distance of that content with
respect to all the other items/objects and users:
Complexity is an O(MC+MU)
M*C distances to be estimated in real time/per day
M*U distances to be estimated in real time/per day
If M= 1 Million
Thus: 10^12 estimations of 10ms, thus 10^10s,
2*317 years !!!
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
37
38. Technologies for Recommendations
G
Objective:
To provide targeted elements on the basis of the elements
descriptors
G
Technical solutions
create distance matrices and matching via direct distance or
similarities estimations, very unfeasible for millions of elements
would be too expensive
making queries on the basis of element profile to get the most
similar. For millions or elements with several aspects or
dimensions in descriptors would be very complex
use some clustering to create group of elements, also based
on distances or similarities. If the groups are too many, the
precisions can be low while the costs are contained.
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
38
39. Similarity Distances
Recipient of the suggestions
Suggested
element
User
Users
Contents
Groups
Content
(played by a user)
Group
(leader or
members)
D(U(s,d);U(s,d))
--no sense--
D(U(s,d);G(s,d))
D(C(s);U(s,d))
D(C(s);C(s))
D(C(s);G(s,d))
D(G(s,d);U(s,d))
D(G(s,d);C(s))
--no sense--
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
39
40. General Distances Models
G
Weighted Models:
T1
T2
i 1
i 1
D(U1; U1) k s xiSd i (U 1, U 2) k d yi Dd i (U1, U2),
G
Vector weighted models:
K s ( x1 Sd1 (U1,U 2), x2 Sd 2 (U1, U 2),, xn SdTs (U1, U 2)) ,
D(U1; U2)
K d ( y1 Dd1 (U1,U 2), y 2 Dd 2 (U1, U 2),, y n DTd (U1,U 2))
G
G
The weights can be defined according to the SN goals.
They can be determined by using multi-linear regressions
techniques.
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
40
42. Clustering among descriptors
G
I C1
K-Means clustering
Based on a multidimensional
distance model among each other
Define the number of clusters
Estimation process to maximize
the cohesion among clusters
I C2
G
Some items can be spare
They are classified in any case
I C3
I C4
G
G
Millions of content items,
thousands of clusters, …
Periodic re-clustering taking into
account all the
content/objects/users
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
42
43. Clustering among descriptors
G
Millions of content items,
ONLY thousands of clusters
G
At each New Object
G
Distance of the new object
with respect to cluster
Centers
Reduction of complexity
I New
G
Usable on recommendations:
UU, UO, OO, etc.
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
43
44. Clustering k-means
G
Good performance in terms of scalability;
G
G
discovery of clusters with arbitrary shape;
ability to deal with noise and outliers;
insensitivity to order of input records;
support for high dimensionality.
Complexity of an O(NKI), where N is the number of elements,
K the number of clusters and I the number of iterations.
k-means has demonstrated the best performances when N is
largely bigger than K and I (Everitt, Landau, Leese, 2001).
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
44
45. K-means problems
G
G
dependency on the availability of numerical absolute
distance estimations between two numerical values
Unfortunatly elements descriptors are
mainly symbolic and in some cases with multiple
values,
coming from both the semantics and concepts they
describe.
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
45
46. K-medoids clustering
G
G
K-medoids adopts as a center of the cluster the element
which has the minimal average (or the median) distance
among the others involved in the cluster.
This means that the complexity is grounded on O(K(NK)2), that for N>>K is an O(N2).
N are the elements
K are the medoids/clusters
G
initially the clusters centers are some selected elements
(Xui & Wunsch, 2009).
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
46
49. Architettura del sistema
I ECLAP
Storage
I P.F
I Utenti
sul
portale
I S.L.I.M.
I Connessione
I Ricezione
fra utenti
suggerimenti
I Partecipazione
al sondaggio
I Us.Ter
I Survey
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
Analysis
49
50. La validazione
I suggerimenti proposti nel sondaggio sono un
sottoinsieme di quelli elaborati dal sistema e
presentano una serie di informazioni relative
agli utenti.
Viene chiesto di votare quanto un
suggerimento è ritenuto interessante.
In questo modo non si valida la qualità
delle metriche ma l’efficacia che hanno i
suggerimenti sulla base delle informazioni
che vengono fornite.
I parametri relativi alla generazione di una raccomandazione sono tutti tracciati e tramite i valori
delle metriche e il voto lasciato dagli utenti si stima l’efficacia che ha mostrare un dettaglio
relativo ad un utente o meno.
Questo è possibile solo perché i dati sulla similarità sono stati calcolati tramite le
procedure precedenti.
È possibile indagare quali sono le tipologie di raccomandazione che vengono gradite
maggiormente.
L’analisi dei dati viene effettuata tramite regressione multilineare per ottenere un modello nella
, forma
, ∙
∙
, ∙
∙
, ∙
∙
I
,
,
∙
∙
∙
,
∙
∙
,
∙
∙
∙
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
50
51. La validazione
5
4
3
1,6
2
1
0
-1
Incidenza sul voto
3,2
1,3
0,3
0,1
-0,4
contenuti
fruiti
Statistica della regressione
R multiplo
0, 9624
F - Value
131,7795
Significatività di F
2,3389E-33
categorie
di
interesse
età
lingua
località
gruppi
Voto <
3
9%
Voto <
3
30%
Voto ≥
3
70%
I Tipologia
Serendipity
Voto ≥
3
91%
I Tipologia
Competenze
Gruppi di appartenza
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
Strategici
Popolarità
51
52. Struttura del Seminario
G
G
G
G
G
G
G
G
Social Networking and knowledge
Semantic and Social Networks
Recommendations and Suggestions
Natural Language Processing System
Knowledge Representation System
Reasoning System
Sistema OSIM
Smart Cities
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
52
53. Natural Language Processing NLP (1)
G
Scenario / Requisiti
Dotare l’IA delle abilità linguistica proprie dell’essere umano
Comprensione e generazione del testo
Contesto multi-language: differenti regole e strutture a seconda della lingua
G
Applicazioni
Generalizzazione delle query nei motori di ricerca
- “Chi si occupa di sistemi distribuiti nell’Università di Firenze ?”
Supporto automatizzato per Help-Desk
Tutoring assistito (e-tutoring, e-teaching…)
Summarization: creare compendi da una collezione
eterogenea di documenti
Machine translation: tradurre testi in lingue diverse
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
53
54. Ambiguità dei linguaggi naturali (1)
G
Scenario / Requisiti
I linguaggi naturali sono ambigui.
Le ambiguità si possono avere a 4 livelli:
Ambiguità lessicale: «attacco» (verbo, sostantivo)
Ambiguità strutturale: «Ieri ho visto l’uomo col telescopio»
«Una vecchia legge la regola»
Ambiguità semantica: «acuto» (persona intelligente, tipo di suono)
Ambiguità pragmatica: «se Buffon non gioca contro la Spagna, l’Italia
perderà»
L’intensione comunicativa viene recepita diversamente dagli interlocutori:
• interpretazione emotiva: l’assenza di Buffon è psicologicamente
fondamentale per i tifosi
• Interpretazione referenziale: l’Italia senza Buffon è più debole
Ciò rende il processo di
elaborazione del linguaggio
naturale molto complicato
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
54
55. Fasi dell’elaborazione in Linguaggio Naturale (1)
Morphological Analysis: le parole vengono analizzate (distinzione dei
morfemi che le compongono) ed i simboli (punteggiature) vengono separati
dalle parole .
Syntactic Analysis: Le sequenze di parole sono trasformate in strutture che
mostrano come le parole sono in relazione l’una con l’altra.
Semantic Analysis: Viene assegnato un significato alle strutture sintattiche
trovate.
Discourse integration: il significato di una frase spesso dipende dalla
frase che la precede e può influenzare quello della frase che la segue.
Pragmatic Analysis: la frase è reinterpretata per determinare il significato
specifico della frase stessa.
“la porta è aperta” necessita di conoscere quale è stata l’intenzione dell’interlocutore:
• Si è creata una corrente d’aria…
• Invito ad entrare liberamente…
• Richiesta affinché qualcuno chiuda la porta…
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
55
56. Fasi dell’elaborazione in Linguaggio Naturale (2)
Testo in linguaggio
naturale
Analisi sintattica
Tokenizer
Analisi morfologica
Part-of-Speech tagger
Chunker
Segmenta il testo in parole e altre sequenze
significative di caratteri (token) ed assegna
una categoria grammaticale ad ogni token
Riconosce suffissi, prefissi e composti
lessicali
Annota ogni parola di una frase con la sua
natura grammaticale all’interno del periodo
(predicato, complemento oggetto, soggetto,
altri complementi ecc….)
Usa le annotazioni precedenti per
raggruppare le frasi nominali e verbali di
ogni sentenza
Produce l’albero sintattico di ogni frase
(syntactic parse tree)
Parser sintattico
Analisi semantica
Le frasi in linguaggio naturale
vengono codificate in una
rappresentazione semantica
(formalismi logici, reti semantiche,
ontologie)
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
56
57. Machine Learning
I sistemi di NLP usano principalmente algoritmi di machine learning addestrati
su corpus di testi annotati a mano
Machine learning
Algorithms
Corpus di testi
annotati a mano
NLP
Testi non
annotati
Testi annotati dal
sistema
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
57
58. Struttura del Seminario
G
G
G
G
G
G
G
G
Social Networking and knowledge
Semantic and Social Networks
Recommendations and Suggestions
Natural Language Processing System
Knowledge Representation System
Reasoning System
Sistema OSIM
Smart Cities
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
58
59. Forme di conoscenza
Implicita
Posseduta dalle persone
Comunicabile in forma verbale o scritta
Tacita
Presente nelle menti degli individui
Difficile da comunicare verbalmente (importante è l’esperienza
sensoriale)
Esplicita
Strutturata (data base, XML+DTD, XML+Shema, ecc.)
Semi-strutturata (XML, ecc.)
Debolmente strutturata (HTML, testi tabulati, ecc.)
Non strutturata (documenti in linguaggio naturale)
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
59
60. Acquisizione e Conservazione
Fonti di Conoscenza
Esperienza diretta
Interazione del soggetto con il suo ambiente
Ragionamento
Deduttivo/inferenza (conclusioni
premesse)
Abduttivo (possibili cause
effetti osservati)
Induttivo (regole generali
fatti specifici)
Comunicazione
Uso di sistemi di segni (in particolare il linguaggio naturale) per
trasferire informazioni da un soggetto a un altro.
Funzione della memoria
Capacità di Conservare nel tempo elementi di conoscenza e
soprattutto di reperirli con efficienza quando occorre farne uso.
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
60
61. La Logica simbolica
Problema
Rappresentare la conoscenza in formato machinereadable (un computer può “leggere” tale
conoscenza rappresentata e utilizzarla per eseguire
compiti d’interesse applicativo)
Soluzione
Rappresentazione dichiarativa tramite logica
simbolica (formale), ed in particolare la logica dei
predicati del primo ordine (first order logic, FOL)
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
61
62. First Order Logic
In FOL tutte le rappresentazioni riguardino un insieme
non vuoto di individui detto universo (o dominio).
Di questi individui possiamo rappresentare proprietà
oppure relazioni che li leghino fra loro. Un fatto è dato
dal sussistere:
di una proprietà di un determinato individuo (es. “Barbara è
bionda”, “Luigi ha 21 anni”)
oppure di una relazione fra più individui (es. “Alberto è più alto
di Barbara”, “Alberto ha dato il suo cellulare a Barbara”, ecc.)
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
62
63. Linguaggi di rappresentazione
Un linguaggio per la rappresentazione di conoscenze è un
linguaggio formale, con sintassi testuale o grafica, le cui
espressioni sono utilizzate per rappresentare elementi di
conoscenza.
Esempio: rappresentare il significato del termine “madre” come
“donna con almeno un figlio”.
Linguaggio naturale:
(x è una madre) se e solo se (x è una donna ed esiste
almeno un y tale che x è genitore di y)
First Order Logic (FOL):
∀ x (MADRE(x) ↔ DONNA(x) ∧ ∃ y GenDi(x,y))
Logic Programming (LP):
madre(X) :- donna(X), genDi(X,Y).
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
63
64. La deduzione
Nel contesto in cui ci stiamo muovendo, per “ragionamento”
s’intende il ragionamento deduttivo (o deduzione)
Una deduzione è un processo che fa passare da alcune
espressioni (dette premesse o ipotesi) a un’espressione (detta
conclusione o tesi), in modo tale da conservare l’eventuale verità
delle premesse: in altre parole, se le premesse sono vere, lo sarà
anche la conclusione.
Ad esempio, dati come premesse
2.
la definizione di “madre”
il fatto che laura è una DONNA
3.
il fatto che laura è GenitoreDi di franco
1.
si può dedurre come conclusione che
laura è una MADRE
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
64
65. Le logiche descrittive
I sistemi di questo tipo hanno preso il nome di logiche
descrittive (description logic, DL). Le DL utilizzano una
sintassi semplificata rispetto a FOL.
Ad esempio, le tre premesse
∀ x (MADRE(x) ↔ DONNA(x) ∧ ∃ y GenDi(x,y))
2.
DONNA(laura)
3.
GenDi(laura,franco)
in logica descrittiva verrebbero rappresentate come
1.
MADRE ≡ DONNA ⊓ ∃GenDi
2.
DONNA(laura)
3.
GenDi(laura,franco)
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
65
66. Risorse
I Tutte
le cose descritte con espressioni RDF
vengono dette Risorse.
I Una
risorsa può essere:
I un'intera
pagina Web (http://www.pippo.it/pluto.html)
I una parte di una pagina Web
I un'intera collezione di pagine
(un sito Web)
I un oggetto non direttamente accessibile via Web (un
libro stampato)
I Le
risorse sono sempre definite da URI
I Qualsiasi
cosa può avere associato un URI
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
66
67. OSIM Ontology
G
Http://openmind.disit.org
G
L’ontologia di dominio di OSIM è composta da 4 ontologie diverse
Academy Life Ontology (Unifi) modella l’ateneo fiorentino in termini di docenti,
corsi, strutture di affiliazione, facoltà, gruppi di ricerca, laboratori, ecc…
Friend of a Friend (FOAF) modella le persone in termini di professori, ricercatori,
phd e relazioni tipo nome, indirizzo, e-mail, settore scientifico, relazioni di
conoscenza, di co-autore di pubblicazioni, ecc …..
Simple Knowledge Organization System (SKOS) che modella ed organizza
semanticamente le competenze delle persone e dei corsi.
Time Ontology (TIME) che modella I concetti di intervalli ed istanti temporali per
quantificare temporalmente i fatti asseriti nell’ontologia.
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
67
68. Time Ontology
I Si
specifica il valore di un
istante ad una certa granularità:
I - Ora
I - Giorno
I - Mese
I - Anno
I- …
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
68
70. OSIM – Frammento Frammento di unifi
Ontologia di dominio
ontology
Relazioni gerarchiche
Relazioni semantiche
uni:PhDStudent
uni:hasCompetence
uni:Research
er
uni:ResearchFellowShip
Frammento di FOAF
uni:AssociatedProfess
or
uni:FullResearcher
foaf:Person
foaf:Organizatio
n
uni:Teache
r
uni:Professor
uni:Coordinator
uni:hasSchoolingCourse
uni:haAffineLivello2
uni:Laboratory
uni:SSD
uni:PhDGradutationCour
se
uni:GraduationCour
se
uni:activeInYear
uni:Master
uni:inSchoolingHasCourseAtTime
uni:takeCourseAtTime
uni:areaCunDi
uni:Department
uni:Course
uni:haAreaCun
uni:Bas
e
uni:schoolingHasCourse
uni:CareerAtTime
uni:Schoolin
g
uni:hasDepartment
uni:Cente
r
uni:haAffineLivello1
uni:schoolingActivity
uni:InterDepartmentAre
a
uni:belongsToArea
uni:hasLaboratory
uni:toBeBasedTo
uni:hasSchoolingCourseInstant
uni:affiliatedFaculty
uni:Facult
y
uni:AreaCUN
uni:hasSchoolingCourseAtTime
time:DurationDescription
time:Instant
Frammento di
SKOS
uni:inSchoolingHasCourseInstant
uni:instantPerson
Year, Week, Second,
Month,
Minute, Hour, Day,
dayOfWeek, DayOfYear,
timeZone
time:inXSDDateTime
time:inDateTime
Frammento di
Time Ontology
skos:Occ
Years, Months, Days, Weeks,
Hours, Minutes, Seconds
time:TemporalEntity
xsd:dateTime
time:TemporalUnit
y
skos:narrower
time:Interval
time:unitType
time:DateTimeDescription
blankNode
skos:type
uni:takeCourseInstant
skos:broader
skos:Concept
time:ProperInterval
time:hasDateTimeInterval
time:DateTimeInterval
skos:ConceptSchema
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
70
71. Struttura del Seminario
G
G
G
G
G
G
G
G
Social Networking and knowledge
Semantic and Social Networks
Recommendations and Suggestions
Natural Language Processing System
Knowledge Representation System
Reasoning System
Sistema OSIM
Smart Cities
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
72
72. Tipi di ragionamento
I Compito
di ragionamento (reasoning task)
è caratterizzato dal tipo di enunciati che si desidera
dedurre da una base di conoscenze
I Procedura
di ragionamento
l’algoritmo che consente la deduzione degli enunciati
I Servizio
di ragionamento
un servizio effettivamente implementato da uno
strumento e messo a disposizione delle applicazioni
che accedono alla base di conoscenze.
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
73
73. Interrogare la conoscenza (3)
I Built-in
I E’
SPARQL
possibile effettuare l’unione di più graph paths
tramite la clausola UNION
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
74
74. Struttura del Seminario
G
G
G
G
G
G
G
G
Social Networking and knowledge
Semantic and Social Networks
Recommendations and Suggestions
Natural Language Processing System
Knowledge Representation System
Reasoning System
Sistema OSIM
Smart Cities
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
75
82. Relazioni verbali (1)
G
G
Il sistema permette di estrarre relazioni semantiche non tassonomiche dalle pagine
dell'Universita’ di Firenze che riguardano i corsi e le persone
Per estrarre le relazioni semantiche dai testi, il sistema analizza le frasi, genera un
grafo a partire dall'albero delle dipendenze di ogni frase e individua eventuali
PATTERN PREDEFINITI presenti sul grafo
I Esempio:
Il portale e-clap è stato sviluppato dal DISIT
SUBJ
DET
AUX
Il
portale e-clap
subj
COMP
AUX
è
stato
sviluppato
è
stato
sviluppato
PREP
dal
DISIT
prep
comp
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
83
84. Validazione (1)
G
OSIM è stato validato e confrontato con Marsilius
La performance dei due sistemi, su un sottoinsieme di 8 dipartimenti, è stata
misurata e confrontata
G
Per sistemi di IR viene tipicamente utilizzato lo standard di validazione TREC (Text
REtrieval Conference)
#(relevant items retrieved)
#(retrieved items)
#(relevant items retrieved)
#(relevant items)
G
G
I documenti / risultati sono considerati rilevanti se soddisfano il tipo di informazione
richiesta, non solo perché contengono tutte le keywords immesse nella ricerca.
Query di validazione eseguite da esperti dei vari domini di conoscenza analizzati
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
85
85. Validazione (2)
G
G
Set di query su un sottoinsieme di 4 dipartimenti
Profondità dei risultati fissata N = 20
Curva Precision – Recall ottenuta con il software standard Trec_Eval
Trec_Eval Precision / Recall Evaluation
OSIM Prec vs Rec
Marsilius Prec vs Rec
1
I Ottimo
ideale
0.8
Precision
G
0.6
0.4
0.2
0
0,1
0,2
0,3
0,4
0,5
0,6
0,7
0,8
0,9
1
Recall
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
86
86. Struttura del Seminario
G
G
G
G
G
G
G
G
Social Networking and knowledge
Semantic and Social Networks
Recommendations and Suggestions
Natural Language Processing System
Knowledge Representation System
Reasoning System
Sistema OSIM
Smart Cities
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
87
87. Department of Information Engineering (DINFO)
DISIT, Distributed Data Intelligence and Technologies
Distributed Systems and Internet Technologies
http://www.disit.dinfo.unifi.it
Main Smart City and OD Projects
• Linked Open Graph: http://log.disit.org
• Sii‐Mobility http://www.sii‐mobility.org
• SmartCityOntology Coll@bora
• SACVAR and OSIM
• see them it via http://log.disit.org
• see http://www.disit.org
I Dottorato,
Univ. Firenze, Paolo Nesi 2013-2014
I 88
88. Department of Information Engineering (DINFO)
DISIT, Distributed Data Intelligence and Technologies
Distributed Systems and Internet Technologies
http://www.disit.dinfo.unifi.it
I SmartCityOntology
•
•
•
•
http://www.disit.org/5606
Title: SmartCity Ontology for Service Inference
Duration: 12 months
Objectives: create an ontology that allows to combine all the data
provided by the city of Florence and the Tuscan region:
• 509 OpenData (Municipality of Florence)
• 119 OpenData (Tuscany Region)
• Timetable TPL
• Street Graph
• Punti di interesse
• Real Time Data from traffic sensors
• Real Time Data from parking sensors
• Real Time Data from AVM systems
• Weather Forrecast (consortium Lamma)
Link: http://log.disit.org, http://www.disit.org/5606
I Dottorato,
Univ. Firenze, Paolo Nesi 2013-2014
I 89
89. Department of Information Engineering (DINFO)
DISIT, Distributed Data Intelligence and Technologies
Distributed Systems and Internet Technologies
http://www.disit.dinfo.unifi.it
SmartCity
Ontology
I http://www.disit.org/5606
I Dottorato,
I 90
Univ. Firenze, Paolo Ne
90. Department of Information Engineering (DINFO)
DISIT, Distributed Data Intelligence and Technologies
Distributed Systems and Internet Technologies
http://www.disit.dinfo.unifi.it
I SmartCityOntology
I Dottorato,
… major macroclasses
Univ. Firenze, Paolo Nesi 2013-2014
I 91
91. Department of Information Engineering (DINFO)
DISIT, Distributed Data Intelligence and Technologies
Distributed Systems and Internet Technologies
http://www.disit.dinfo.unifi.it
I Linked
•
•
•
•
•
Open Graph
Title: Linked Open Graph: Social and Linked Open Data
Navigation
Duration: 24 months, derived from EC project ECLAP
Objectives:
• Design and develop tools for graphical navigation on Open
Data and Linked Open Data
Link:
• LOG LOD http://log.disit.org
• Also used in www.ECLAP.eu Social Graph:
http://www.eclap.eu/116088
Examples for: dbPedia, Europeana, British Museum,
LinkedGeo Data, Cultura Italia, Sii-Mobility, ICARO Cloud,
MyStoryPlayer, OSIM Knowledge Modeling and reasoning
I Dottorato,
Univ. Firenze, Paolo Nesi 2013-2014
I 92
92. DISIT Lab, Distributed Data Intelligence and Technologies
Distributed Systems and Internet Technologies
Department of Information Engineering (DINFO)
http://www.disit.dinfo.unifi.it
Linked Open Graph SmartCityOntology
http://log.disit.org
I DISIT
Lab (DINFO UNIFI), Paolo Nesi, 14
Febbraio 2014
I 93
93. DISIT Lab, Distributed Data Intelligence and Technologies
Distributed Systems and Internet Technologies
Department of Information Engineering (DINFO)
http://www.disit.dinfo.unifi.it
Piattaforma di sviluppo (beta)
http://servicemap.disit.org
I DISIT
Lab (DINFO UNIFI), Paolo Nesi, 14
Febbraio 2014
I 94
94. Department of Information Engineering (DINFO)
DISIT, Distributed Data Intelligence and Technologies
Distributed Systems and Internet Technologies
http://www.disit.dinfo.unifi.it
I Linked
Open Graph on dbPedia
I http://log.disit.org
I Dottorato,
Univ. Firenze, Paolo Nesi 2013-2014
I 95
95. Department of Information Engineering (DINFO)
DISIT, Distributed Data Intelligence and Technologies
Distributed Systems and Internet Technologies
http://www.disit.dinfo.unifi.it
Linked Open Graph for LOD
A browsing tool to explore LOD sparql services via their entry point. To explore RDF
elements and view which contents or users are linked with that.
With just one click (or tap) over a node you can see appear the navigation panel that
allows you to:
• Explore/Reduce a node of the graph.
• Focus the visualization over a node.
• Open a specified content and view it's info.
• Direct accessing to the info associated with an entity, attributes and their values.
• Filtering relationships, inverting the filtering.
• Save your linked open graph with your preferences and navigations and get their access
via email, that you can share with your colleagues for reading and further browsing and
change.
A list of check buttons, one for each relation kind, to turn on/off the visualization of
relations from the LOG.
LOG tool is free of use for no profit organizations. You can embed the LOG tool in your web
pages.
I Dottorato,
Univ. Firenze, Paolo Nesi 2013-2014
I 96
96. DISIT Lab, Distributed Data Intelligence and Technologies
Distributed Systems and Internet Technologies
Department of Information Engineering (DINFO)
http://www.disit.dinfo.unifi.it
Sii-Mobility (Smart City nazionale)
• Titolo: Supporto di Interoperabilità Integrato per i Servizi al Cittadino
e alla Pubblica Amministrazione
• Ambito: Trasporti e Mobilità Terrestre
• Obiettivi:
1. ridurre i costi sociali della mobilità
2. semplificare l’uso dei sistemi di mobilità
3. Sviluppo di soluzioni e applicazioni funzionanti e sperimentazione
4. Contribuire al miglioramento degli standard nazionali ed
internazionali
• Coordinatore Scientifico: Paolo Nesi, DISIT DINFO UNIFI
• Partner: ECM; Swarco Mizar; University of Florence (svariati gruppi+CNR); Inventi
In20; Geoin; QuestIT; Softec; T.I.M.E.; LiberoLogico; MIDRA; ATAF; Tiemme; CTT
Nord; BUSITALIA; A.T.A.M.; Sistemi Software Integrati; CHP; Effective Knowledge;
eWings; Argos Engineering; Elfi; Calamai & Agresti; KKT; Project; Negentis.
• Durata: 36 months; Costo: circa 14 Meuro
• Link: http://www.disit.dinfo.unifi.it/siimobility.html
I DISIT
Lab (DINFO UNIFI), Paolo Nesi, 14
Febbraio 2014
I 97
97. DISIT Lab, Distributed Data Intelligence and Technologies
Distributed Systems and Internet Technologies
Department of Information Engineering (DINFO)
http://www.disit.dinfo.unifi.it
• Sperimentazioni principalmente in
Toscana
• Sperimentazioni piu’ complete in
aree primarie ad alta integrazione
dati
• Integrazione con i sistemi presenti
I DISIT
Lab (DINFO UNIFI), Paolo Nesi, 14
Febbraio 2014
I 98
98. DISIT Lab, Distributed Data Intelligence and Technologies
Distributed Systems and Internet Technologies
Department of Information Engineering (DINFO)
http://www.disit.dinfo.unifi.it
Sii‐Mobility: Scenari principali
•
soluzioni di guida/percorso connessa/o
–
•
Piattaforma di partecipazione e sensibilizzazione
–
•
tariffazione dinamica e per categoria di veicoli;
gestione rete condivisa di scambio dati fra servizi (PA e privati)
–
•
Politiche pay‐per‐use, monitoraggio comportamento degli utenti;
gestione dinamica dei confini delle aree a traffico controllato
–
•
contribuzione a standard, verifiche e validazione dei dati, riconciliazione dei dati, etc.;
integrazione di metodi di pagamento e di identificazione
–
•
Politiche di incentivazione e di dissuasione dell’uso del veicolo, Crediti di mobilità, monitoraggio flussi;
interoperabilità ed integrazione dei sistemi di gestione
–
•
per ricevere dal cittadino informazioni, il cittadino come sensore intelligente, informare e formare il cittadino,
tramite totem, applicazioni mobili, web applications, etc.;
gestione personalizzata delle politiche di accesso
–
•
servizi personalizzati, segnalazioni, il veicolo/la persona riceve comandi e informazioni in tempo reale ma
modo personalizzato e contestualizzato;
affidabilità dei dati e separazione delle responsabilità, Integrazione di open data, riconciliazione, ….;
monitoraggio della domanda e dell’offerta di trasporto pubblico in tempo reale
–
soluzioni per l’integrazione e l’elaborazione dei dati.
I DISIT
Lab (DINFO UNIFI), Paolo Nesi, 14
Febbraio 2014
I 99
99. DISIT Lab, Distributed Data Intelligence and Technologies
Distributed Systems and Internet Technologies
Department of Information Engineering (DINFO)
http://www.disit.dinfo.unifi.it
Architettura Sii‐Mobility
Sistemi di Gestione
Open Data:
ambientali, energia,
ordinanze, grafo, ….
Social Media:
twitter FaceBook, Blogs
Acq Dati Emegenza
Supporto: bigliettazione,
prenotazione, pianificazione, …
Orari, percorsi, fermate,
Valori attuali, politiche di
azione, etc.
Data ingestion and
pre processing
Sensori e
Apparati di
Rilevazione
Supporto: data analytics,
semantic computing
profiles,
modelli di
costo, users,
..
Data Processing:
validation, integration,
geomap.., reconcil, …
Additional
algorithms, ..…..
Big Data processing grid
Interfaccia di Controllo e
Monitoraggio
Infomobilità e publicazione
Piattaforma di partecipazione e di
Sensibilizzazione
Info. Services for
Totem, Mobile, ..
Gestori Flotte, AVM,
TPL, Parcheggi,
Autostrade, Car
Sharing, Bike
sharing, ZTL,
direzionatori …..
Centrale Operativa:
Supporto alle decisioni,
simulazioni, …
API Data publication verso
altri centri di Servizi o Utenti
Gestori:
Agrgegazione dati,
propagazione azioni
Kit veicoli e Attuatori
Supporto di Interoperabilita’ Integrato
Applicazioni
Web e Totem
Applicazioni
Mobili
API altri Sii‐Mobility
API altre Centrali SC
API Centrali nazionali
Interoperabilità da e verso altri
sistemi di gestione integrata e SC
Dati integrati dal territorio
Infrastruttura HW Sii‐Mobility
I DISIT
Lab (DINFO UNIFI), Paolo Nesi, 14
Febbraio 2014
I 100
100. DISIT Lab, Distributed Data Intelligence and Technologies
Distributed Systems and Internet Technologies
Department of Information Engineering (DINFO)
http://www.disit.dinfo.unifi.it
Problematiche di Ricerca “Smart”
• Modellazione della conoscenza con semantica coerente per
effettuare deduzioni corrette sfruttando informazioni numeriche e
simboliche, su grandi volumi e flussi di dati, automazione
– Problematiche derivate da oltre 1000 OD e PD, servizi con modelli diversi
e formati diversi: resilienza, qualità, misura, accesso, integrazione real
time, …
– Tecniche di: modellazione, semantic computing, scheduling, …
• Ricerca su integrazione e modellazione dati:
– Alto livello: predizione su servizi e comportamenti, correlazioni
inattese, situazioni critiche, flusso dei viaggiatori, …
– Problemi e algoritmi di riconciliazione del dato, tracciamento e
versionamento, reputazione, filtraggio, integrazione OD e LOD
internazionali/nazionali, validazione e verifica formale, …. …. ….
I DISIT
Lab (DINFO UNIFI), Paolo Nesi, 14
Febbraio 2014
I 101
101. DISIT Lab, Distributed Data Intelligence and Technologies
Distributed Systems and Internet Technologies
Department of Information Engineering (DINFO)
http://www.disit.dinfo.unifi.it
Riconciliazione servizi
Service_Key
Road_Key
R_address
extendName
Cod_
toponimo
BELLAVISTALargo_F.lli_Alinari_15
048017LargoF.lliAlinari
Largo F.lli Alinari LARGO FRATELLI ALINARI
RT04801701866TO
CASA_DEL_LAGOLungarno__A._Vespucci_58
048017LungarnoA.Vespucci
Lungarno A.
Vespucci
LUNGARNO AMERIGO
VESPUCCI
RT04801701874TO
COSMOPOLITANVia_F._Baracca187
048017ViaF.Baracca
Via F. Baracca
VIA FRANCESCO BARACCA RT04801702987TO
SAN_PAOLO_IMIVIA DE'
048017VIADE'VECCHIETTI
VIA_DE'_VECCHIETTI_22/R
VECCHIETTI
VIA DEI VECCHIETTI
RT04801702383TO
CREDITO_ARTIGIANOVIA_DE'_BONI_1
048017VIADE'BONI
VIA DE‘ BONI
VIA DEI BONI
RT04801702326TO
Auditorium_al_DuomoVia_De'_Cerretani_54/r
048017ViaDe'Cerretani
Via De' Cerretani VIA DEI CERRETANI
SiiMobility:isIn
Service
RT04801702317TO
owl:sameAs
Nuovo
elemento Road
I DISIT
Elemento Road
grafo stradale
Lab (DINFO UNIFI), Paolo Nesi, 14
Febbraio 2014
I 102
102. DISIT Lab, Distributed Data Intelligence and Technologies
Distributed Systems and Internet Technologies
Department of Information Engineering (DINFO)
http://www.disit.dinfo.unifi.it
Architettura del sistema
I DISIT
Lab (DINFO UNIFI), Paolo Nesi, 14
Febbraio 2014
I 103
103. DISIT Lab, Distributed Data Intelligence and Technologies
Distributed Systems and Internet Technologies
Department of Information Engineering (DINFO)
http://www.disit.dinfo.unifi.it
Ingestion (un esempio)
I DISIT
Lab (DINFO UNIFI), Paolo Nesi, 14
Febbraio 2014
I 104
104. DISIT Lab, Distributed Data Intelligence and Technologies
Distributed Systems and Internet Technologies
Department of Information Engineering (DINFO)
http://www.disit.dinfo.unifi.it
Validazione e Verifica
Query in linguaggio SPARQL
Query di verifica del collegamento tra gli oggetti creati e
quelli del grafo stradale
• Si seleziona un elemento del grafo stradale e si richiedono tutti gli elementi
ad esso collegati
• Nel caso di grandi quantità di dati risulta necessario contare il numero di
elementi collegati
Oggetto
Totale
Riconciliabili (7 comuni
di Firenze)
Riconciliati
Previsioni meteo
286
7
7
Statistiche del comune
115
115
115
Uffici Pubblici
752
176
176
Servizi
28560
3559
3502
Statistiche sulle vie di Firenze 7987
7987
I DISIT
7987
Lab (DINFO UNIFI), Paolo Nesi, 14
Febbraio 2014
I 105
105. Department of Information Engineering (DINFO)
DISIT, Distributed Data Intelligence and Technologies
Distributed Systems and Internet Technologies
http://www.disit.dinfo.unifi.it
I Coll@bora
•
•
•
•
I
•
Title: Collaborative Support for Parents and Operators of Disabled
Duration: 24 months
Cost: 1 Meuro
Objectives: providing strong advantages for
1.
Relatives interested in facilitating relations with the
management team;
2.
Associations in order to offer a better service to the families
and people with disabilities by providing a collaborative
support to the involved teams, but also to manage the wealth
of knowledge, to support the training of the staff, etc.
Coll@bora provides a secure collaboration tool for the teams
and for the association to support the families and the disabled
people.
Link: http://www.disit.dinfo.unifi.it/collabora.html
I Dottorato,
Univ. Firenze, Paolo Nesi 2013-2014
I 106
106. Department of Information Engineering (DINFO)
DISIT, Distributed Data Intelligence and Technologies
Distributed Systems and Internet Technologies
http://www.disit.dinfo.unifi.it
I Coll@bora
•
•
•
Objective 1: Study and development of a
platform for collaboration and management for
team (consisting of parents, family members,
physicians, physician assistants, volunteers,
etc..) for disable support in privacy
Objective 2: Study and development of web
applications, and mobile smartTv to support
the activities of assistance and service
Objective 3: validation of the solution
I Dottorato,
Univ. Firenze, Paolo Nesi 2013-2014
I 107
107. Department of Information Engineering (DINFO)
DISIT, Distributed Data Intelligence and Technologies
Distributed Systems and Internet Technologies
http://www.disit.dinfo.unifi.it
I Coll@bora
I Dottorato,
Univ. Firenze, Paolo Nesi 2013-2014
I 108
108. Department of Information Engineering (DINFO)
DISIT, Distributed Data Intelligence and Technologies
Distributed Systems and Internet Technologies
http://www.disit.dinfo.unifi.it
SACVAR/OSIM: Open Mind Innovative Space
• Title: Open Mind Innovative Space, and SACVAR
• Duration: 36 months, 12 as OSIM e 24 as SACVAR
• Objectives:
1. Knowledge mining construction from people competence
2. Semantic search engine on mined knowledge
3. Web pages and Blog mining and analysis in Natural Language
4. Match making and affective computing analysis
– We are using these tools for social media analysis, to analyze citizens
appreciation and comments on Smart City services.
• Link: http://openmind.disit.org
•
•
http://www.disit.dinfo.unifi.it/osim.html
http://www.disit.dinfo.unifi.it/sacvar.html
I Dottorato,
Univ. Firenze, Paolo Nesi 2013-2014
I 109
124. ECLAP
Metadata
Ingestion
Server
AXCP back office services
O
A
I
Ingestion and Harvesting
P
M
H
ECLAP Social
Service Portal
Resource Injection
Content Analysis
Archive
partner
Archive
partner
Archive
partner
Content
Retrieval
Library
Library
partner
Library
partner
Networking
IPR Wizard/CAS
Metadata Editor
Content Processing
Content Aggregation and Play
Semantic Computing and Sugg.
Content Indexing and Search
Content Upload Management
Metadata
Metadata
Export
Database +
semantic database
Content Upload
Social Network
connections
E-Learning
Support
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
I 125
125. Estrazione e Deduzione della
Conoscenza via Modelli Semantici:
From Social Network to Smart City
seminario per il Corso di Dottorato 2014
Prof. Paolo Nesi
Dipartimento di Ingegneria dell’Informazione
University of Florence
Via S. Marta 3, 50139, Firenze, Italy
tel: +39-055-4796523, fax: +39-055-4796363
DISIT Lab
Paolo.nesi@unifi.it
http://www.disit.dinfo.unifi.it/
Dottorato, Univ. Firenze, Paolo Nesi 2013-2014
126