Projects A.A. 2012/2013

Main topics for projects and th ithesis

Progetto di Ingegneria Progetto di Ingegneria Informatica

I diti i ti l P tt di I i • I crediti associati al Progetto di Ingegneria Informatica prevedono attività progettuale e di sperimentazione che possono messere svolte sotto la guida di possono messere svolte sotto la guida di tutti i docenti di Ingegneria Informatica

• Sono associabili ai crediti della tesi per gli t d ti d ll’ di t 509studenti dell’ordinamento 509

• I progetti sono illustrati sul sito http://pii.dei.polimi.it/

• Le docenti Antola e Tanca sono incaricate di registrare tutti i progetti (suddivisione alfabetica)( )

• Di seguito illustriamo i progetti del nostro gruppo, non tutti ancora presenti sul sito

• Questi lucidi verranno pubblicati anche Questi lucidi verranno pubblicati anche sulla pagina del corso di BD

BasicsBasics• A number of ongoing projects• There is room:

for design/implementation or– for design/implementation  or – for more theoretical research, or – for experiments

• For first level (bachelor) students: seeFor first level  (bachelor) students: see http://pii.dei.polimi.it/

• For second level (master) students: projects can be• For second level  (master) students:  projects can be performed as master thesis projects (tesina) or, within the same research area, as research master thesis (tesi)research area, as research master thesis (tesi)

Information Overload and Noise:Personalization and context awarenessInformation Overload and Noise:Personalization and context awarenessPersonalization and context‐awarenessPersonalization and context‐awareness

• Context‐based personalization: shaping answers (to queries) according to h ’ f d i i (i )the user’s preferences and situation (i.e., context).– model and collect characteristics of the users (or groups of)– mostly implicit (behavioral analysis sensing )mostly implicit (behavioral analysis, sensing, …)– non‐functional (e.g., data quality)

user processes

user situation

user processes• dynamic• context and evolution of

user profile

• static• context‐based

preference based• involve sensing• context learning

evolution ofpersonalization

• static• user‐based

context learning issues

Information Personalization:Context aware data tailoringInformation Personalization:Context aware data tailoringContext‐aware data tailoringContext‐aware data tailoring

• context aware• observables• context schema

Context Modelling

• instantiation• validation• reasoning

Context Sensing

• context‐aware data

• context‐aware operations

Context‐Aware 

Behaviour

design-time run-time

Context representationContext representation 

Th ti f t t d li ti h i d l t• The notions of context and personalization have acquired a lot of attention in the last years; context‐aware systems are pervading everyday lifepervading everyday life.

• The system uses the knowledge of the current user context to focus on the data that are really usefuly

• Contextual  and personal aspects need to be modeled so that the system can use them for information tailoring and focus

6

ProjectsContacts: {tanca,quintare,garza,mazuran, panigati,rabosio,rauseo}@elet.polimi.it

• Context-aware services and APIs

• Context sharing and the evolution between different contextrepresentations

C t t f d d ti• Context-aware preferences and recommendations

• Contextual non-functional properties (e.g. data quality)

All the above projects propose design and

implementation or experiments

More theoretical:

• Logic-based frameworks for contextual systems

Projects (continued)Contacts:

{tanca,schreiber,panigati,mazuran@elet.polimi.it

• GREEN MOVE: electric car-sharing• In this project we have designed the whole

d t b d th t tdatabase and the context-awarenesslayer. Possible evolutions on the data mining issues

• SENSORI: monitoring energy consumption in a neighborhood• In this project we need to design the p j g

service ontology and the context-awareness layer

TreeRuler projectAssociation

DBrules

Association rule mining

QQuery Intensional answer

1) Mine frequent association rules from a global database2) Store mined rules3) Learn from frequent rules “common” properties of co-3) Learn from frequent rules common properties of co

related information4) Use the inferred knowledge to provide faster and

summarized (intensional) answers to user’s queries. S i d l f l f fi i i

9

Summarized answers are also useful for refining queries (and answers).

Projects

Contacts: {tanca,quintare,rauseo}@elet.polimi.it{ ,q , }@ p

• Using the system within new languagef ( )frameworks (e.g. RDF)

• Optimizations• Database scouting on the web for• Database scouting on the web for

providing data for experimenting the systemy

• Data abstraction for schema induction• Application to finding regularities in

bi l i l d tbiological data

Context-aware Wireless Sensor Context aware Wireless Sensor Networks

• PerLa: a middleware and a context-aware language to query WSN (seeaware language to query WSN (seeprof. Schreiber’s web site)

• Extension of the language towards full Extension of the language towards full distribution

• Usage of this language for monitoringg g g gvarious phenomena: energy-awareness, cars’ and production

t ’ b h i tsystems’ behaviours, etc.• Applied also in Green Move and

SensoriSensori

ProjectsMany more ongoing project proposals