Stream Reasoning with LARS Harald Beck Minh Dao-Tran Thomas Eiter - - PowerPoint PPT Presentation

stream reasoning with lars
SMART_READER_LITE
LIVE PREVIEW

Stream Reasoning with LARS Harald Beck Minh Dao-Tran Thomas Eiter - - PowerPoint PPT Presentation

Stream Reasoning with LARS Harald Beck Minh Dao-Tran Thomas Eiter The 2nd Stream Reasoning Workshop December 8-9, 2016 Outline LARS Overview (Revised) LARS Syntax and Semantics Recent and Ongoing LARS Research @ KBS LARS Collaboration


slide-1
SLIDE 1

Stream Reasoning with LARS

Harald Beck Minh Dao-Tran Thomas Eiter The 2nd Stream Reasoning Workshop December 8-9, 2016

slide-2
SLIDE 2

Outline

LARS Overview (Revised) LARS Syntax and Semantics Recent and Ongoing LARS Research @ KBS LARS’ Collaboration

slide-3
SLIDE 3

LARS Overview

LARS is a logical framework with a rule-based language that

  • ffers

◮ generic window operators, ◮ different ways to refer to time, and ◮ an ASP-like semantics

for (analyzing) stream reasoning.

slide-4
SLIDE 4

LARS Setting: Streams

11 12 13 14 15 16 17 18 19 20 21 22 23

  • a

p p e a r e d ( , t r

2 5

)

  • a

p p e a r e d ( 1 , t r

2 5

)

  • a

p p e a r e d ( 1 , b

2 7 A

)

Data Streams D = (T, υ) T = [0, 50] υ =    12 → {appeared(0, tr 25)}, 17 → {appeared(1, tr 25)}, 19 → {appeared(1, b27A)}   

slide-5
SLIDE 5

LARS Setting: Streams

11 12 13 14 15 16 17 18 19 20 21 22 23

  • a

p p e a r e d ( , t r

2 5

)

  • a

p p e a r e d ( 1 , t r

2 5

)

  • a

p p e a r e d ( 1 , b

2 7 A

)

  • e

x p ( 2 , t r

2 5

)

  • e

x p ( 2 , b

2 7 A

)

Interpretation Stream S⋆ = (T ⋆, υ⋆) ⊇ D T ⋆ = [0, 50] υ⋆ =    12 → {appeared(0, tr 25)}, 21 → {exp(2, tr 25)}, 17 → {appeared(1, tr 25)}, 23 → {exp(2, b27A)}, 19 → {appeared(1, b27A)}   

slide-6
SLIDE 6

LARS Window Functions and Operators

11 12 13 14 15 16 17 18 19 20 21 22 23

  • a

p p e a r e d ( , t r

2 5

)

  • a

p p e a r e d ( 1 , t r

2 5

)

  • a

p p e a r e d ( 1 , b

2 7 A

)

Window functions S′ = (T ′, υ′) = w(S, t) T ′ = [15, 20] υ′ = {17 → {appeared(1, tr 25)}, 19 → {appeared(1, b27A)}} Window operators ⊞w

slide-7
SLIDE 7

(Revised) LARS Syntax

α ::=

slide-8
SLIDE 8

(Revised) LARS Syntax

α ::= a | ¬α | α∧α | α∨α | α → α

◮ standard logical operators

slide-9
SLIDE 9

(Revised) LARS Syntax

α ::= a | ¬α | α∧α | α∨α | α → α | ♦α | α | @tα

◮ standard logical operators ◮ various ways for time references

slide-10
SLIDE 10

(Revised) LARS Syntax

α ::= a | ¬α | α∧α | α∨α | α → α | ♦α | α | @tα | ⊞wα

◮ standard logical operators ◮ various ways for time references ◮ window operators which can be nested

⊞60 ⊞10 ♦appeared(s, b1)

slide-11
SLIDE 11

(Revised) LARS Syntax

α ::= a | ¬α | α∧α | α∨α | α → α | ♦α | α | @tα | ⊞wα | ⊲α

◮ standard logical operators ◮ various ways for time references ◮ window operators which can be nested

⊞60 ⊞10 ♦appeared(s, b1)

◮ Reset operator

slide-12
SLIDE 12

(Revised) LARS Syntax

α ::= a | ¬α | α∧α | α∨α | α → α | ♦α | α | @tα | ⊞wα | ⊲α

◮ standard logical operators ◮ various ways for time references ◮ window operators which can be nested

⊞60 ⊞10 ♦appeared(s, b1)

◮ Reset operator ◮ Rules

@T+Lexp(M, V ) ← ⊞5@Tappeared(N, V ), plan(N, M, V , L).

slide-13
SLIDE 13

(Revised) LARS Entailment

Structure M = S⋆, W , B

slide-14
SLIDE 14

(Revised) LARS Entailment

Structure M = S⋆, W , B Substream S = (T, υ) of S⋆: currently considered window Time point t ∈ T

slide-15
SLIDE 15

(Revised) LARS Entailment

Structure M = S⋆, W , B Substream S = (T, υ) of S⋆: currently considered window Time point t ∈ T M, S, t a iff a ∈ υ(t) or a ∈ B ,

slide-16
SLIDE 16

(Revised) LARS Entailment

Structure M = S⋆, W , B Substream S = (T, υ) of S⋆: currently considered window Time point t ∈ T M, S, t a iff a ∈ υ(t) or a ∈ B , M, S, t ¬α iff M, S, t α, M, S, t α ∧ β iff M, S, t α and M, S, t β, M, S, t α ∨ β iff M, S, t α or M, S, t β, M, S, t α → β iff M, S, t α or M, S, t β,

slide-17
SLIDE 17

(Revised) LARS Entailment

Structure M = S⋆, W , B Substream S = (T, υ) of S⋆: currently considered window Time point t ∈ T M, S, t a iff a ∈ υ(t) or a ∈ B , M, S, t ¬α iff M, S, t α, M, S, t α ∧ β iff M, S, t α and M, S, t β, M, S, t α ∨ β iff M, S, t α or M, S, t β, M, S, t α → β iff M, S, t α or M, S, t β, M, S, t ♦α iff M, S, t′ α for some t′ ∈ T,

slide-18
SLIDE 18

(Revised) LARS Entailment

Structure M = S⋆, W , B Substream S = (T, υ) of S⋆: currently considered window Time point t ∈ T M, S, t a iff a ∈ υ(t) or a ∈ B , M, S, t ¬α iff M, S, t α, M, S, t α ∧ β iff M, S, t α and M, S, t β, M, S, t α ∨ β iff M, S, t α or M, S, t β, M, S, t α → β iff M, S, t α or M, S, t β, M, S, t ♦α iff M, S, t′ α for some t′ ∈ T, M, S, t α iff M, S, t′ α for all t′ ∈ T ,

slide-19
SLIDE 19

(Revised) LARS Entailment

Structure M = S⋆, W , B Substream S = (T, υ) of S⋆: currently considered window Time point t ∈ T M, S, t a iff a ∈ υ(t) or a ∈ B , M, S, t ¬α iff M, S, t α, M, S, t α ∧ β iff M, S, t α and M, S, t β, M, S, t α ∨ β iff M, S, t α or M, S, t β, M, S, t α → β iff M, S, t α or M, S, t β, M, S, t ♦α iff M, S, t′ α for some t′ ∈ T, M, S, t α iff M, S, t′ α for all t′ ∈ T , M, S, t @t′α iff M, S, t′ α and t′ ∈ T ,

slide-20
SLIDE 20

(Revised) LARS Entailment

Structure M = S⋆, W , B Substream S = (T, υ) of S⋆: currently considered window Time point t ∈ T M, S, t a iff a ∈ υ(t) or a ∈ B , M, S, t ¬α iff M, S, t α, M, S, t α ∧ β iff M, S, t α and M, S, t β, M, S, t α ∨ β iff M, S, t α or M, S, t β, M, S, t α → β iff M, S, t α or M, S, t β, M, S, t ♦α iff M, S, t′ α for some t′ ∈ T, M, S, t α iff M, S, t′ α for all t′ ∈ T , M, S, t @t′α iff M, S, t′ α and t′ ∈ T , M, S, t ⊞wα iff M, S′, t α where S′ = w(S, t) ,

slide-21
SLIDE 21

(Revised) LARS Entailment

Structure M = S⋆, W , B Substream S = (T, υ) of S⋆: currently considered window Time point t ∈ T M, S, t a iff a ∈ υ(t) or a ∈ B , M, S, t ¬α iff M, S, t α, M, S, t α ∧ β iff M, S, t α and M, S, t β, M, S, t α ∨ β iff M, S, t α or M, S, t β, M, S, t α → β iff M, S, t α or M, S, t β, M, S, t ♦α iff M, S, t′ α for some t′ ∈ T, M, S, t α iff M, S, t′ α for all t′ ∈ T , M, S, t @t′α iff M, S, t′ α and t′ ∈ T , M, S, t ⊞wα iff M, S′, t α where S′ = w(S, t) , M, S, t ⊲α iff M, S⋆, t α ,

slide-22
SLIDE 22

Scenario

U1 Kagran 25 27A Kagraner Platz 26 Carminweg

slide-23
SLIDE 23

LARS Programs

11 12 13 14 15 16 17 18 19 20 21 22 23

  • a

p p e a r e d ( , t r

2 5

)

  • a

p p e a r e d ( 1 , t r

2 5

)

  • a

p p e a r e d ( 1 , b

2 7 A

)

@T+Lexp(M, V ) ← ⊞5@Tappeared(N, V ), plan(N, M, V , L). takeBus(N) ← ⊞+2♦ exp(N, B), bus(B), not takeTram(N). takeTram(N) ← ⊞+5♦ exp(N, Tr), tram(Tr), not takeBus(N).

slide-24
SLIDE 24

LARS Programs

11 12 13 14 15 16 17 18 19 20 21 22 23

  • a

p p e a r e d ( , t r

2 5

)

  • a

p p e a r e d ( 1 , t r

2 5

)

  • a

p p e a r e d ( 1 , b

2 7 A

)

  • e

x p ( 2 , t r

2 5

)

  • e

x p ( 2 , b

2 7 A

)

@T+Lexp(M, V ) ← ⊞5@Tappeared(N, V ), plan(N, M, V , L). takeBus(N) ← ⊞+2♦ exp(N, B), bus(B), not takeTram(N). takeTram(N) ← ⊞+5♦ exp(N, Tr), tram(Tr), not takeBus(N).

slide-25
SLIDE 25

LARS Programs

11 12 13 14 15 16 17 18 19 20 21 22 23

  • a

p p e a r e d ( , t r

2 5

)

  • a

p p e a r e d ( 1 , t r

2 5

)

  • a

p p e a r e d ( 1 , b

2 7 A

)

  • e

x p ( 2 , t r

2 5

)

  • e

x p ( 2 , b

2 7 A

)

  • t

a k e B u s ( 2 )

@T+Lexp(M, V ) ← ⊞5@Tappeared(N, V ), plan(N, M, V , L). takeBus(N) ← ⊞+2♦ exp(N, B), bus(B), not takeTram(N). takeTram(N) ← ⊞+5♦ exp(N, Tr), tram(Tr), not takeBus(N).

slide-26
SLIDE 26

LARS Programs

11 12 13 14 15 16 17 18 19 20 21 22 23

  • a

p p e a r e d ( , t r

2 5

)

  • a

p p e a r e d ( 1 , t r

2 5

)

  • a

p p e a r e d ( 1 , b

2 7 A

)

  • e

x p ( 2 , t r

2 5

)

  • e

x p ( 2 , b

2 7 A

)

  • t

a k e T r a m ( 2 )

@T+Lexp(M, V ) ← ⊞5@Tappeared(N, V ), plan(N, M, V , L). takeBus(N) ← ⊞+2♦ exp(N, B), bus(B), not takeTram(N). takeTram(N) ← ⊞+5♦ exp(N, Tr), tram(Tr), not takeBus(N).

slide-27
SLIDE 27

Recent and Ongoing LARS Research @ KBS

Answer Update for Rule-based Stream Reasoning [Beck et al., 2015] Contrasting RDF Stream Processing Semantics [Dao-Tran et al., 2015] Equivalence of LARS programs [Beck et al., 2016b] Ongoing implementation based on Truth Maintenance Systems

slide-28
SLIDE 28

LARS’ Collaboration

Vrije Universiteit Amsterdam (Jacopo Urbani and Hamid Bazoobabdi)

◮ Dipper engine

Alpen-Adria-University Klagenfurt (Konstantin Schekotihin)

◮ Rule-based Stream Reasoning for Intelligent

Administration of Content-Centric Networks [Beck et al., 2016a]

◮ LotTraveller

slide-29
SLIDE 29

References I

Harald Beck, Minh Dao-Tran, and Thomas Eiter. Answer Update for Rule-based Stream Reasoning. In 24th International Joint Conference on Artificial Intelligence (IJCAI), July 25-31, 2015, Buenos Aires, Argentinia, 2015. Harald Beck, Bruno Bierbaumer, Minh Dao-Tran, Thomas Eiter, Hermann Hellwagner, and Konstantin Schekotihin. Rule-based Stream Reasoning for Intelligent Administration of Content-Centric Networks. In JELIA, pages 522–528, 2016.

slide-30
SLIDE 30

References II

Harald Beck, Minh Dao-Tran, and Thomas Eiter. Equivalent stream reasoning programs. In 24th International Joint Conference on Artificial Intelligence (IJCAI), July 9-15, 2016, New York, USA, 2016. Minh Dao-Tran, Harald Beck, and Thomas Eiter. Contrasting RDF Stream Processing Semantics. In The 5th Joint International Semantic Technology Conference, JIST 2015, Yichang, China, November 11-13, 2015.