Querying ontology-based knowledge bases equipped with non-monotonic rules through a case study within the framework of Cultural Heritage. It focuses on 3D underwater surveys on the Xlendi wreck which is represented by an OWL2 knowledge base with a large dataset. The paper aims at improving the interactions between the archaeologists and the knowledge base providing new queries that involve non-monotonic rules in order to perform qualitative spatial reasoning.
To this end, the knowledge base initially represented in OWL2-QL is translated into an equivalent Answer Set Programming (ASP) program and is enriched with a set of non-monotonic ASP rules suitable to express default and exceptions. An ASP query answering approach is proposed and implemented.
Furthermore due to the increased expressiveness of non-monotonic rules it provides spatial reasoning and spatial relations between artifacts query answering which is not possible with query answering languages such as SPARQL and SQWRL.
Query answering with non monotonic rules a case study of archaeology qualitative spatial reasoning
1. Query Answering With Non-Monotonic
Rules: A Case Study of Archaeology
Qualitative Spatial Reasoning
Mohamed Ben Ellefi, Pierre Drap, Laurent Garcia, Fabien Garreau,
Claire Lefèvre, Odile Papini, Igor Stéphan and Eric Würbel
GCAI 2019
5th Global Conference on Artificial Intelligence
17-19 September 2019 - Bolzano, Italy
2. Xlendi shipwreck
• datable to approximately 7th century BC at
a depth of 110 m.
• one of the best-preserved archaeological
sites in Malta datable to the early
Phoenician period.
• the oldest archaeological shipwreck in the
Mediterranean
The framework of the GROPLAN project (http://www.lsis.org/groplan)
Context: Xlendi Shipwreck
3. ● The site was surveyed by photogrammetry at 110
meters depth.
“Photogrammetry is the science of
making 3D measurements from
photographs...”
Context: Xlendi Shipwreck
4. The semantic of the cargo:
➢ The nature of the ship ?
➢ The ship destination ?
➢ The origin ?
(…)
Distribution AnalysisOrientations AnalysisTypology Analysis
Representing the dataset in a knowledge base (KB)
How to manage and produce knowledge ?
10. Dilemma: get closer to Archaelogists reasoning ...
➢ Archaeologists do not perform any computation.
➢ In contrast, they focus on the relationships
(expressed in natural languages) between artifacts to
perform qualitative spatial reasoning.
➢ The reasoning is only based on symbolic and spatial
qualitative information without calculation.
11. Visibility/Closeness Relation
Visibility between two artifacts means that there is no obstacle between them.
NW
W
SW SE
S
N
E
NE
Archaeologists focus on closeness (direct neighbor) between artifacts, through the visibility relation:
12. Non-Monotonic Rules for Spatial Reasoning
Let A, B and C three artifacts.
If wof (A,B) and
● for any artifact C such that wof (A,C)
● for any artifact C such that
nwof(A,C), and not neof(B,C) nor eof(B,C), and
● for any artifact C such that
swof(B,C), and not eof(B,C) nor seof(B,C),
B
and nor seof(B,C), and
nor eof(B,C),and not neof(B,C),
C
C
C
A
→ then B is visible (the closest) from A
towards the West direction.
West
East
13. From OWL2 KB to ASP
[1] Baget, J., Garcia, L., Garreau, F., Lef`evre, C., Rocher, S., Stéphan, I.: Bringing existential variables in answer set
programming and bringing non-monotony in existential rules: two sides of the same coin. Ann. Math.Artif. Intell.
82(1-3), 3–41 (2018)
1. OWL2 KB translated into the existential rules formalism via a set of tools
graal (https://graphik-team.github.io/graal).
1. Resulting set of existential rules translated into ∃ASP.
1. ∃ASP program translated into a standard ASP program by skolemization of existential variables.
All tools [1] available on ASPIQ project Website: http://aspiq.lsis.org
14. From OWL2 KB To non-monotonic rules for query
answering
15. Implementation
Xlendi ontological knowledge base:
● TBox : 483 axioms about 69 classes and 124 properties (32
object properties and 92 data type properties).
● ABox: 6210 facts describing 75 amphorae and 55 grinding
stones (in the process of growth ++),
○ 16770 additional facts describing the cardinal relations
between the 130 artifacts.
ASP query answering demonstrator
★ a Web solution that combines
○ Web JS version of Clingo solver (https://potassco.org/)
○ 2D Web drawing of Xlendi orthophoto (powered by
Raphaël JS: http://dmitrybaranovskiy.github.io/raphael).
DEMO: http://www.arpenteur.org/xlendi_asp
16. DEMO: Query 1
➢ What is the name of amphorae with height between 0.4
meter and 0.6 meter ?
○ ans(Y) :- amphorae(X), hasName(X,Y),
hasHeight(X,Z), Z < 600, Z > 400.
Solving...
Answer: 1
ans("Amphore_A47") ans("Amphore_A93") ans("Amphore_A23") ans("Amphore_A27")
ans("Amphore_A39") ans("Amphore_A01") ans("Amphore_A05") ans("Amphore_A31")
ans("Amphore_A65") ans("Amphore_A75") ans("Amphore_A77") ans("Amphore_A69")
ans("Amphore_A02") ans("Amphore_A06") ans("Amphore_A71") ans("Amphore_A60")
ans("Amphore_A21") ans("Amphore_A56") ans("Amphore_A92") ans("Amphore_A66")
ans("Amphore_A18") ans("Amphore_A63") ans("Amphore_A30") ans("Amphore_A70")
ans("Amphore_A64") ans("Amphore_A28") ans("Amphore_A79") ans("Amphore_A16")
ans("Amphore_A55") ans("Amphore_A61") ans("Amphore_A89") ans("Amphore_A24")
ans("Amphore_A43") ans("Amphore_A90") ans("Amphore_A76") ans("Amphore_A52")
ans("Amphore_A42") ans("Amphore_A11") ans("Amphore_A74") ans("Amphore_A03")
ans("Amphore_A07") ans("Amphore_A62")
SATISFIABLE
Models : 1
Calls : 1
Time : 2.341s
17. DEMO: Query 2
➢ Which amphorae are of type "Pithecusse_366" and are
visible from amphora Amphore_A31 ?
ans(Y) :- amphorae("Amphore_A21"), amphorae(Y),
visible("Amphore_A32", Y),
hasTypologyName(Y,"Pithecusse_366").
Solving...
Answer: 1
ans("Amphore_A30")
SATISFIABLE
Models : 1
Calls : 1
Time : 2.279s
18. DEMO: Query 2
➢ Which amphorae are of type "Pithecusse_366" and are
NOT visible from amphora Amphore_A31 ?
○ ans(Y) :- amphorae("Amphore_A31"), amphorae(Y),
not visible("Amphore_A31", Y),
hasTypologyName(Y,"Pithecusse_366").
Solving...
Answer: 1
ans("Amphore_A14") ans("Amphore_A25") ans("Amphore_A04") ans("Amphore_A07")
ans("Amphore_A08") ans("Amphore_A34") ans("Amphore_A28")
SATISFIABLE
Models : 1
Calls : 1
Time : 2.281s
19. Conclusion
➢ A POC for ASP query answering approach over an OWL2-QL KB equipped with non-monotonic ASP rules.
➢ The increased expressiveness of non-monotonic rules allows for querying spatial
relations like closeness between artifacts
○ This is not possible with languages such as SPARQL and SQWRL.
➢ ASP query answering demonstrator is provided online comining
○ a Web version of Clingo solver
○ a 2D Web drawing of Xlendi shipwreck orthophoto.
Up
Down
Future Directions:
★ The extension of cardinal relations to 3D spatial relations in order to query a 3D
representation of the cargo which is closer to reality.
★ The future of surveyed data have to be in sharing, exchanging, reusing the
analysed data :
○ Linked Open Data
○ Easy query GUI for Users