Prof. Carlo Tasso University of Udine Artificial Intelligence Laboratory Group BIRLA Hyderabad, 8-10/11/2004

Transcription

1 Intelligent Digital Platforms for the semantic Web: New Technologies for Accessing and Filtering Information and Knowledge Prof. Carlo Tasso University of Udine Artificial Intelligence Laboratory infofactory Group BIRLA Hyderabad, 8-10/11/2004

2 Contents infofactory The General Disciplinary Area Problems Possible Solutions Resaerch Activities The infofactory idea Intelligent Platforms: Experiments, Prototypes, Systems Open Research Problems Actions in the Project EU-INDIA cross cultural dissemination Demo 09/11/2004 2

3 infofactory Area of Research Information & Communication Technologies Internet Tools for accessing E-Contents: Web Digital Information & Services (current shift from information to services) The tools: Artificial Intelligence techniques: linguistic/semantic analysis, user modeling Web technologies 09/11/2004 3

4 infofactory General Problems Information Explosion E-Contents Overload Strategic Value of Information & Knowledge (IST) Information is the vehicle of Knowledge 09/11/2004 4

5 Information on the Web From Information Producers to Information Consumers infofactory ideas/concepts/events/ authors, information producers unstructured info/ free text documents/multimedia docs./ /audio-video/ E-Contents WEB/repositories/ data banks search (pull) intermediaries delivery (push) consumers 09/11/2004 5

6 Problems with Information (and Knowledge) Problems in searching, accessing, delivering Two perspectives: PULL & PUSH infofactory Pull: Oversupply, overload Miss-retrieval (common accuracy of serach engines: 20%-30%) Untimeliness Push: Miss-delivery Oversupply Waste Untimeliness Two difficult linguistic phenomena: Polisemy, Sinonimity 09/11/2004 6

7 infofactory The fundamental limitations of current approaches and tools Key-word based: NO semantic analysis of texts One-size-fits all approach: all users are treated the same way, NO comprehension of individual information needs and preferences 09/11/2004 7

8 infofactory Towards a Solution 09/11/2004 8

9

10 Automatic Personalization (in Information Access) infofactory Personalization means delivering: the right content to the right user in the right moment and in the most suitable way 09/11/

11 infofactory Adaptive Personalization Adaptivity allows to automatically tailor system behaviour and processing to the specific characteristics of the (observed) user behaviour Machine Learning, Conceptual Linguistic Analysis, Knowledge Representation, User Modeling, Intelligent Agents Semantic WEB 09/11/

12 Semantic Personalized Information Filtering infofactory Filtering information by means of a semantic analysis of a document and by matching it against an individual user profile (Brajnik G., Guida G. and Tasso C. User Modeling in Expert Man-Machine Interfaces: A Case Study in Intelligent Information Retrieval. IEEE Transaction on Systems, Man, and Cybernetics, 20(1), 1990, pp ) 09/11/

13 The IFT Algorithm (Minio & Tasso, 1996) infofactory Positive Sample Documents Negative Sample Documents Documents to be filtered (html, xml, pdf, postscript, doc, text, latex, ) PROFILE BUILDER LINGUISTIC PROCESSOR User Profile Conceptual Content of a Document Relevance feedback from the user MATCHING Evaluation of the relevance of the document (da C.Tasso, P.Omero, La Personalizzazione dei Contenuti WEB, F.Angeli, Milano, 2002.) 09/11/

14 Experimental Evaluation of the IFT Algorithm: Precision infofactory (Asnicar, Di Fant & Tasso, 1997) 09/11/

15 Experimental Evaluation infofactory of the IFT Algorithm: : Ranking Quality Search Engines IFT (Asnicar, Di Fant & Tasso, 1997) 09/11/

16 The TIPS Project in FP5 IST (500 k ) infofactory Innovative advanced services in electronic publishing (scholarly journals) Filtering service on new incoming information Application in the field of High Energy Physics (Brajnik G., Mizzaro S., Tasso C., Venuti F., Strategic Help in User Interfaces for Information Retrieval, Journal of the American Society for Information Science and Technology 53(5), 2002, pp ) (Mizzaro and Tasso, 2002) tips.sissa.it 09/11/

17 infofactory The TORII Portal 09/11/

18

19 The infofactory Concept

20 infofactory The infofactory Concept Delivering personalized information services starting from any digital information source (E- Content) Adaptive personalization based on content-based individual profiles Overcoming information overload by filtering, categorizing, ranking, alerting, and monitoring digital sources Delivering the right content, to the right user, at the most adequate time, and in the most suitable way 09/11/

21 the infofactory Digital E-Content Space infofactory REPOSITORIES REPOSITORIES (data banks, OPAC, CD-Rom,...) INTERNET INTERNET (web sites, portals, news, directories, e-newpapers, e- mail, web TV/radio,webcams,...) DIGITAL DIGITAL TV TV (digital TV broadcast, satellite TV channels, interactive TV, streaming services, videoconferences,...) E-Content Consumer Space text/data, text /data, 2D/3D 2D/3D graphics, graphics digital digital audio/video, audio/video, (any format, any standard) B2B - B2B2C Personalized Personalized Information Information Services Services B2C (individual end users, virtual communities, associations, ) (info producers, information brokers, info re-disributors, service&content providers, vertical portals, infomediators, web clippers, ) - source discovery - content identification - content retrieval - content filtering - content classification - content ranking - content re-mixing - c. recommendation - content reformulation - content monitoring - content delivery - content push - alerting - user profiling - conceptual analysis on PC, PDA, mobile devices (wap, gprs, umts,...), TV, playstation,... 09/11/

22 infofactory The infofactory Technical Approach Cognitive (content-based) Filtering Analysis of the conceptual content of textual documents Semantic networks for describing user interests Filtering obtained by matching documents against profiles Adaptive Personalization Individual user profiles Machine learning algorithms for automatically building and updating user profiles on the basis of positive (negative) examples 09/11/

23 infofactory Intelligent Platforms 09/11/

24 Contruction of personal Data Banks Web Monitoring and Clipping service Finding Experts Match macking module between expert competences and requests filtering Automatic categorization Antispam Filter infofactory ifmonitor ifexpert ifmail Others.. Cognitive filtering Conceptual analysis of natural language Multichannel Opp. Alerting Adaptive Personalization Classification Intelligent Crawling infofactory technologies and basic components 09/11/

25 Contruction of personal Data Banks Web Monitoring and Clipping service Finding Experts Match macking module between expert competences and requests filtering Automatic categorization Antispam Filter infofactory Application Level ifmonitor ifexpert ifmail Others.. System Level Cognitive filtering Conceptual analysis of natural language Multichannel Alerting Adaptive Personalization Classification Intelligent Crawling infofactory technologies and basic components PLATFORM Level 09/11/

26 Intelligent Platform1: ifmonitor for Web Monitoring and Filtering

27 infofactory Filtering Web Information Using the IFT algorithm for filtering information gathered from the Web by means of an opportunistic (automatic) human like navigation (crawling), relevant for a given user profile (Tasso and Armellini, 1999) 09/11/

28 infofactory Monitoring Web Information Periodically analysing the content of Web sites and portals for timely discovering new relevant information 09/11/

29 ifmonitor Services infofactory Personalized Filtering Services based on conceptual analisys of textual Web documents: Automatic construction of personalized THEMATIC DATA BANKS of Web documents WEB MONITORING, continuous periodic control of online sources and multichannel ALERTING 09/11/

30 infofactory What are they useful for? Technological monitoring Monitoring scientific and technical journals Reputation management Press review Competitive intelligence Keeping knowledge worker always updated Crisis management 09/11/

31 infofactory What are they useful for? -2 Monitoring events Monitoring financial funding open calls Discovering research groups belonging to other organizations in the same area of interests or research 09/11/

32 infofactory What advantages? Reducing manual activities for searching information (higher productivity 24/24, 7/7 and 365/365) Finding information published in sites not (or rarely) indexed by standard search engines Timely discovery and multi-channel delivery of new relevant information Much more precise than other competitors (90% vs % of standard search engines) Innovative information services (personalized services, alerting, mobile access) 09/11/

33 UTENTE (Addetto Comunicazione, Ufficio Stampa, DG, ) Documenti dell utente Inserimento documenti dal propriopc* Inserimento manuale di notizie tramite Scanner o da Web* Utility (ordinamenti, cartelle, pdf, mail, annotazione*, opt-in*) * Servizi Opzionali Argomenti di interesse, esempi Consultazione e ricerca tramite Motore ifportal Statistiche sui media Suddivisione in sottocategorie ed inoltro* ifmonitor profili ifclassify Costruzione iniziale BD WEB nuove notizie Monitoraggio Periodico Banca Dati Tematica Personalizzata di doc. Web (che cresce incrementalmente nel tempo) Pubblicazione Automatica* (previa validazione opt-in) Dicono di noi Portale Istituzionale Sorgenti Online

34 infofactory ifportal: a portal for accessing ifmonitor services 09/11/

35

36

37

38

39

40

41

42

43

44 Mobile access infofactory 09/11/

45 infofactory Intelligent Platform 2: ifexpert for Expert Finding 09/11/

46 infofactory Representing human expertise and searching for it Using the same approach of the IFT algorithm for analysing text and matching against profiles Competence profiles are built (and later updated) by analysing textual information describing personal expertise, projects, papers, and so on A premium service of the MACSI-net Web site: matchmaking between requests for experts and an expert data base An innovative service in the Portale TIROCINI (Stage Portal), a portal for managing stages in industries for university students: match-making between stage requests and student competence & preferences. 09/11/

47 infofactory a WEB site exploiting the ifexpert platform 09/11/

48 infofactory 09/11/

49 infofactory Portale TIROCINI: a WEB portal exploiting the ifexpert match-making capabilities 09/11/

50 More than Students, with their Competences, credits, preferences

51 More than 1000 Companies and Public Agencies

52 More than Students, with their Competences, credits, Prefrences (structured information)

53 More than Students, with their Competences, credits, Prefrences (unstructured information)

54 ifexpert automatically identifies the best matching projects, taking into account unstructured information

55 infofactory Intelligent Platform 3: ifmail for Mail Filtering and Antispam 09/11/

56 Filtering and Classification of e Messages infofactory Filtering incoming messages, discarding the not interesting ones (spam) Automatically categorizing messages into user-defined categories of specific interest Category profiles are learned automatically, by observing user s behaviour Identifying the most relevant messages and alerting on Mobile devices 09/11/

57

58 ifmail Architecture infofactory 09/11/

59 ifmail Architecture - 2 infofactory Modulo WebMail Database delle rappresentazioni interne (MySQL) nuove rappresentazioni interne rappresentazioni interne rappresentazione interna Modulo Costruttore delle rappresentazioni interne (Java) richiesta o costruzione rappr. interna Modulo Frame (Java) richiesta di categorizzazione feedback risposte Modulo Categorizzatore (Java) richieste feedback e categorizzazione risposte Modulo Agente di Monitoraggio (Java) messaggi categorizzati Modulo Alerting Multicanale Modulo IFT (Java) messaggi interessanti (notifica) Dispositivi mobili 09/11/

60 Evaluation of ifmail infofactory F1 COLLEZIONE A (SPAM) Saturation level above 0.8 F1 (per sessione) F1 (per sessione) F1 (consiglio) F1 (consiglio) 1 1 0,9 0,9 0,8 0,8 0,7 0,7 Microaverage F1 Microaverage F1 0,6 0,6 0,5 0,5 0,4 0,4 0,3 0,3 0,2 0,2 0,1 0, categorizzate categorizzate Learning times AntiSpam Filter 106 Categorization errors 7 09/11/

61 infofactory Intelligent Platform 4: mytwm E-Learning Platform for delivering value-added added services to University students 09/11/

62 Personalization in the TWM Portal for value- added services to University students infofactory The TWM portal provides several services for University students Several services are supporting learning activities The student can personalize the portal and access only to the preferred services, and associate alerting Fully XML-SCORM compliant 09/11/

63 infofactory 09/11/

64 infofactory 09/11/

65 infofactory ifforum: an advanced forum, with personalization and tools for anaysing social interaction 09/11/

66 Standard Forum Capabilities

67 Message exchange network

68 Centrality measure: discovering leaders

69 PCA and clique analysis: discovering groups

70 Discovering users that are information bridges

71 Exploiting personalization and filtering for proactive recommendation

72 infofactory Open Research Problems New Alghorithms for automatic adaptation, filtering, learning Extension of current filtering approaches to multimedia information Exploitation of new media (mobile) Personalization in Knowledge Management and e- Learning systems 09/11/

73 infofactory Actions in our Project EU EU-India for Cross Cultural Dissemination A new digital (E-Contents) platform integrating tools coming fom mytwm, ifexpert, Portale Tirocini Data banks and Web Monitoring and Alerting in the areas of interest Personalization, information sharing and advanced communication tools Match-making tools for cross country matching in specific areas 09/11/

74 infofactory retrieving filtering classifying. support learning sharing info and know. opening relationships reducing distances abbreviating times 09/11/

75 infofactory tç Ç à tà äx Éy à{x TÜà y v tä \ÇàxÄÄ zxçvx _tuéütàéüç? hç äxüá àç Éy hw Çx? ÑÜÉyA VtÜÄÉ gtááé 09/11/