Evaluation Techniques for Rule-Based Deduction in Dynamic Databases - - PowerPoint PPT Presentation

evaluation techniques for rule based deduction in dynamic
SMART_READER_LITE
LIVE PREVIEW

Evaluation Techniques for Rule-Based Deduction in Dynamic Databases - - PowerPoint PPT Presentation

www.imdea.org Evaluation Techniques for Rule-Based Deduction in Dynamic Databases Joaqun Arias IMDEA Software Institute madrid institute for advanced studies in software development technologies www.imdea.org Joaqun Arias Herrero


slide-1
SLIDE 1

www.imdea.org

Evaluation Techniques for Rule-Based Deduction in Dynamic Databases

Joaquín Arias

IMDEA Software Institute

madrid institute for advanced studies in software development technologies

slide-2
SLIDE 2

www.imdea.org

Joaquín Arias Herrero

http://www.software.imdea.org/~joaquin.arias 2011-2014 BSc Computer Science (UPM) Thesis “Memory management in logic programming with tabling” 2014-2015 MSc Software and Systems (UPM) Thesis ‘‘Towards a Generic Interface to Integrate CLP and Tabled Execution (Modular TCLP)” 2013-today Research at the IMDEA Software Institute Advisor: Manuel Carro Liñares PhD topic: “Evaluation Techniques for Rule-Based Deduction in Dynamic Database”

madrid institute for advanced studies in software development technologies

slide-3
SLIDE 3

www.imdea.org

Big Data

Volume Variety Velocity Veracity

madrid institute for advanced studies in software development technologies

slide-4
SLIDE 4

www.imdea.org

Big Data vs Stream/Dynamic Data

Volume Variety Velocity Veracity

madrid institute for advanced studies in software development technologies

slide-5
SLIDE 5

www.imdea.org

Big Data vs Stream/Dynamic Data

Volume Variety Velocity Veracity Dynamic Data are generated by: Real-time sensors System events Social media messages ...

madrid institute for advanced studies in software development technologies

slide-6
SLIDE 6

www.imdea.org

Big Data vs Stream/Dynamic Data

Volume Variety Velocity Veracity Dynamic Data are generated by: Real-time sensors System events Social media messages ... They could be analyzed and used to: Optimize retailers’ stock Optimize supply chain or delivery route Detect fraudulent transactions Detect cyber attacks Optimize traffic flows ...

madrid institute for advanced studies in software development technologies

slide-7
SLIDE 7

www.imdea.org

Example: Rail ticket retail

madrid institute for advanced studies in software development technologies

slide-8
SLIDE 8

www.imdea.org

Example: Rail ticket retail

Rule If there is great activity in the social media regarding a concert, reduce the price-offers and increase the frequencies of trains with the city where the concert will take place.

madrid institute for advanced studies in software development technologies

slide-9
SLIDE 9

www.imdea.org

Example: Rail ticket retail

Rule If there is great activity in the social media regarding a concert, reduce the price-offers and increase the frequencies of trains with the city where the concert will take place. Event Great activity is detected in Malaga regarding a concert in Alicante.

madrid institute for advanced studies in software development technologies

slide-10
SLIDE 10

www.imdea.org

Example: Rail ticket retail

Rule If there is great activity in the social media regarding a concert, reduce the price-offers and increase the frequencies of trains with the city where the concert will take place. Event Great activity is detected in Malaga regarding a concert in Alicante. Action Modify the offers between Malaga and Alicante.

madrid institute for advanced studies in software development technologies

slide-11
SLIDE 11

www.imdea.org

Example: Rail ticket retail

Rule If there is great activity in the social media regarding a concert, reduce the price-offers and increase the frequencies of trains with the city where the concert will take place. Event Great activity is detected in Malaga regarding a concert in Alicante. Action Modify the offers between Malaga and Alicante. Event The concert is postponed one week / bad weather forecast.

madrid institute for advanced studies in software development technologies

slide-12
SLIDE 12

www.imdea.org

Example: Rail ticket retail

Rule If there is great activity in the social media regarding a concert, reduce the price-offers and increase the frequencies of trains with the city where the concert will take place. Event Great activity is detected in Malaga regarding a concert in Alicante. Action Modify the offers between Malaga and Alicante. Event The concert is postponed one week / bad weather forecast. Action Update the offers (in real time) and report the event.

madrid institute for advanced studies in software development technologies

slide-13
SLIDE 13

www.imdea.org

Our approach:

The application of TCLP , an interface that integrate tabled execution with constraints solvers in a logic programming environment (Prolog) to the evaluation of Dynamic Databases with Rule-Based deduction.

Advantages:

Expressiveness Will be easier to express complex rules and non-trivial reasoning. Readable Constraint logic programming is closer to the know- ledge representation that e.g. map/reduce functions. Non-monotonicity Revision of previous inferences rules due to modified data (by insertion, modification or deletion). Less computation Automatic reuse of inferences brought by tabling. Only one language Rules and actions can be written using the same for- malism.

madrid institute for advanced studies in software development technologies

slide-14
SLIDE 14

www.imdea.org

Our approach:

The application of TCLP , an interface that integrate tabled execution with constraints solvers in a logic programming environment (Prolog) to the evaluation of Dynamic Databases with Rule-Based deduction.

Challenges:

Design new constraint solvers Design and integrate in TCLP a solver to reduce the required accesses to databases. Re-engineer the tabling engine To avoid using obsolete data / results. Optimize TCLP To minimize the recomputation due to modified data by using previously tabled results.

madrid institute for advanced studies in software development technologies

slide-15
SLIDE 15

www.imdea.org

Value

Thanks

madrid institute for advanced studies in software development technologies