SlideShare a Scribd company logo
Looking for Invariant Operators in Argumentation
Stefano Bistarelli1 Francesco Santini1 Carlo Taticchi2
stefano.bistarelli@unipg.it francesco.santini@unipg.it carlo.taticchi@gssi.it
1Universit`a degli Studi di Perugia, Italy
2Gran Sasso Science Institute (GSSI), L’Aquila, Italy
Looking for Invariant Operators in Argumentation
Looking for Invariant Operators in Argumentation
Looking for Invariant Operators in Argumentation
Looking for Invariant Operators in Argumentation
Index
1 Background - Argumentation Frameworks
2 Definition - Invariant Operators
3 Conclusion and Future Work
Carlo Taticchi Looking for Invariant Operators in Argumentation 3 / 14
Argumentation Frameworks1
A human-like fashion to deal with knowledge
Definition (AF)
An Abstract Argumentation Framework is a pair G = A, R where A is a
set of arguments and R is a binary relation on A.
1
PHAN MINH DUNG. On the Acceptability of Arguments and its Fundamental Role in Nonmonotonic Reasoning, Logic
Programming and n-Person Games. Artif. Intell., 77(2):321–358, 1995.
Carlo Taticchi Looking for Invariant Operators in Argumentation 4 / 14
Sets of Extensions
Definition (Conflict-free extensions)
Let G = A, R be an AF. A set E ⊆ A is conflict-free in G if there are no
a, b ∈ A | (a, b) ∈ R.
Carlo Taticchi Looking for Invariant Operators in Argumentation 5 / 14
Sets of Extensions
Definition (Conflict-free extensions)
Let G = A, R be an AF. A set E ⊆ A is conflict-free in G if there are no
a, b ∈ A | (a, b) ∈ R.
Scf (G) = {{}, {1}, {2}, {3}, {4}, {5},
{1, 2}, {1,4}, {1, 5}, {2,5}, {3, 5}, {1, 2, 5}}
Carlo Taticchi Looking for Invariant Operators in Argumentation 5 / 14
Dung’s Semantics
Definition (Admissible Semantics)
Let G = A, R be an AF. A set E ⊆ A is admissible in G if E is
conflict-free and each a ∈ E is defended by E.
Sadm(G) = {{}, {1},{2}, {3}, {4}, {5},
{1, 2},{1,4},{1, 5}, {2, 5}, {3, 5}, {1, 2, 5}}
Carlo Taticchi Looking for Invariant Operators in Argumentation 6 / 14
Dung’s Semantics
Definition (Admissible Semantics)
Let G = A, R be an AF. A set E ⊆ A is admissible in G if E is
conflict-free and each a ∈ E is defended by E.
Similar definitions for complete, grounded, preferred and stable.
Carlo Taticchi Looking for Invariant Operators in Argumentation 6 / 14
Reinstatement Labelling2
Every argument can be labelled in, out or undec.
Example (Labelling of an AF)
2
MARTIN CAMINADA. On the Issue of Reinstatement in Argumentation. Logics in Artificial Intelligence: 10th European
Conference, JELIA 2006 Liverpool, UK, September 13-15, 2006 Proceedings, pages 111–123, 2006.
Carlo Taticchi Looking for Invariant Operators in Argumentation 7 / 14
Robustness3,4
A property of an AF to withstand changes
Tries to answer the following questions:
• Is it possible to change the outcome of a debate according to a
particular semantics or meaning?
• If so, how easy could it be to perform such change?
• And which consequences does it bring?
3
S. BISTARELLI, F. FALOCI, F. SANTINI, AND C. TATICCHI. Robustness in abstract argumentation frameworks. In
Proceedings of the 29th International Florida Artificial Intelligence Research Society Conf. FLAIRS, page 703, 2016.
4
CARLO TATICCHI. A Study of Robustness in Abstract Argumentation Frameworks. Proceedings of the Doctoral
Consortium of AI*IA 2016, Genova, Italy, November 29, 2016, pages 11–16. CEUR-WS.org, 2016.
Carlo Taticchi Looking for Invariant Operators in Argumentation 8 / 14
Modifying Operators
Allow to construct AFs adding an attack relations at time.
Definition (Modifying operator)
Let G = A, R ∈ Fn be an AF. A modifying operator is a function
m : Fn → Fn such that m(G) = A, m(R) , where m(R) ⊇ R.
Carlo Taticchi Looking for Invariant Operators in Argumentation 9 / 14
Modifying Operators
Allow to construct AFs adding an attack relations at time.
Definition (Modifying operator)
Let G = A, R ∈ Fn be an AF. A modifying operator is a function
m : Fn → Fn such that m(G) = A, m(R) , where m(R) ⊇ R.
Carlo Taticchi Looking for Invariant Operators in Argumentation 9 / 14
Modifying Operators
Allow to construct AFs adding an attack relations at time.
Definition (Modifying operator)
Let G = A, R ∈ Fn be an AF. A modifying operator is a function
m : Fn → Fn such that m(G) = A, m(R) , where m(R) ⊇ R.
Carlo Taticchi Looking for Invariant Operators in Argumentation 9 / 14
Comparing AFs
We have defined a partial order on the set Fn of all AFs with n
arguments and all possible combinations of attack relations.
Definition (AFs inclusion w.r.t. attacks)
Let G1 = A, R1 and G2 = A, R2 be two AFs. We say that G1 is
included in G2 w.r.t. attacks if R1 ⊆ R2 and we write G1 ≤A G2.
Figure: G1 Figure: G2
Carlo Taticchi Looking for Invariant Operators in Argumentation 10 / 14
Comparing Semantics
Definition (Semantics inclusion)
Let S and S be two sets of extensions. We say that S ⊆ S if and only if
∀E ∈ S ∃E ∈ S | E ⊆ E .
Figure: Sadm(G1) = {{}, {1}, {1, 3}} Figure: Sadm(G2) = {{}, {1}}
Carlo Taticchi Looking for Invariant Operators in Argumentation 11 / 14
Invariant Operators for Admissible Semantics
Example (AF with 4 arguments)
Figure: Sadm = {{}, {1}, {4}, {1, 4}}
Carlo Taticchi Looking for Invariant Operators in Argumentation 12 / 14
Invariant Operators for Admissible Semantics
Example (AF with 4 arguments)
Figure: Sadm = {{}, {1}, {4}, {1, 4}}
2 → 1: ∃c = b such that there is an odd length sequence of attacks from
c to a, but not from a to c.
Carlo Taticchi Looking for Invariant Operators in Argumentation 12 / 14
Invariant Operators for Admissible Semantics
Example (AF with 4 arguments)
Figure: Sadm = {{}, {1}, {4}, {1, 4}}
2 → 3: there is no odd length sequence of attacks from b to a.
Carlo Taticchi Looking for Invariant Operators in Argumentation 12 / 14
Invariant Operators for Admissible Semantics
Example (AF with 4 arguments)
Figure: Sadm = {{}, {1}, {4}, {1, 4}}
4 → 3: c ∈ in(L) such that (b, c) ∈ R.
Carlo Taticchi Looking for Invariant Operators in Argumentation 12 / 14
Invariant Operators for Admissible Semantics
Example (AF with 4 arguments)
Figure: Sadm = {{}, {1}, {4}, {1, 4}}
Carlo Taticchi Looking for Invariant Operators in Argumentation 12 / 14
Summing Up
Invariant local expansion operators for the admissible semantics allow for:
• adding an attack in an AF, while
• maintaining the admissible semantics unchanged
a → b Condition
in → in never possible
in → out c ∈ in(L) | (b, c) ∈ R
in → undec c ∈ undec(L) | (c, c) /∈ R and (b, c) ∈ R
out → in (b, a) ∈ R or ∃c ∈ out | (c, b) ∈ R
out →
in there is no odd length sequence of attacks from b to a
out → out or ∃c = b | there is an odd length sequence of attacks
undec from c to a, but not from a to c
undec → in never possible
undec → out no condition required
undec → undec no condition required
Carlo Taticchi Looking for Invariant Operators in Argumentation 13 / 14
Further Research
Next possible steps:
• To design invariant operators with respect to the complete,
grounded, preferred, semi-stable, and stable semantics.
• To find the essential extensions for which every change inside
modifies the semantics, while changes outside do not.
• To study local expansion operators also for semiring-based
weighted AFs5.
• To consider different notions of equivalence6 (e.g. local
equivalence) and additional modifications of AFs (as the deletion of
attack or the addition/removal of arguments).
5
S. BISTARELLI, F. ROSSI, AND F. SANTINI. A Collective Defence Against Grouped Attacks for Weighted Abstract
Argumentation Frameworks. FLAIRS Conference 2016: 638-643
6
E. OIKARINEN, AND S. WOLTRAN. Characterizing strong equivalence for argumentation frameworks. 2011. Artificial
Intelligence 175(14-15):1985–2009.
Carlo Taticchi Looking for Invariant Operators in Argumentation 14 / 14
Further Research
Next possible steps:
• To design invariant operators with respect to the complete,
grounded, preferred, semi-stable, and stable semantics.
• To find the essential extensions for which every change inside
modifies the semantics, while changes outside do not.
• To study local expansion operators also for semiring-based
weighted AFs5.
• To consider different notions of equivalence6 (e.g. local
equivalence) and additional modifications of AFs (as the deletion of
attack or the addition/removal of arguments).
5
S. BISTARELLI, F. ROSSI, AND F. SANTINI. A Collective Defence Against Grouped Attacks for Weighted Abstract
Argumentation Frameworks. FLAIRS Conference 2016: 638-643
6
E. OIKARINEN, AND S. WOLTRAN. Characterizing strong equivalence for argumentation frameworks. 2011. Artificial
Intelligence 175(14-15):1985–2009.
Carlo Taticchi Looking for Invariant Operators in Argumentation 14 / 14
Further Research
Next possible steps:
• To design invariant operators with respect to the complete,
grounded, preferred, semi-stable, and stable semantics.
• To find the essential extensions for which every change inside
modifies the semantics, while changes outside do not.
• To study local expansion operators also for semiring-based
weighted AFs5.
• To consider different notions of equivalence6 (e.g. local
equivalence) and additional modifications of AFs (as the deletion of
attack or the addition/removal of arguments).
5
S. BISTARELLI, F. ROSSI, AND F. SANTINI. A Collective Defence Against Grouped Attacks for Weighted Abstract
Argumentation Frameworks. FLAIRS Conference 2016: 638-643
6
E. OIKARINEN, AND S. WOLTRAN. Characterizing strong equivalence for argumentation frameworks. 2011. Artificial
Intelligence 175(14-15):1985–2009.
Carlo Taticchi Looking for Invariant Operators in Argumentation 14 / 14
Further Research
Next possible steps:
• To design invariant operators with respect to the complete,
grounded, preferred, semi-stable, and stable semantics.
• To find the essential extensions for which every change inside
modifies the semantics, while changes outside do not.
• To study local expansion operators also for semiring-based
weighted AFs5.
• To consider different notions of equivalence6 (e.g. local
equivalence) and additional modifications of AFs (as the deletion of
attack or the addition/removal of arguments).
5
S. BISTARELLI, F. ROSSI, AND F. SANTINI. A Collective Defence Against Grouped Attacks for Weighted Abstract
Argumentation Frameworks. FLAIRS Conference 2016: 638-643
6
E. OIKARINEN, AND S. WOLTRAN. Characterizing strong equivalence for argumentation frameworks. 2011. Artificial
Intelligence 175(14-15):1985–2009.
Carlo Taticchi Looking for Invariant Operators in Argumentation 14 / 14
Looking for Invariant Operators in Argumentation
Stefano Bistarelli1 Francesco Santini1 Carlo Taticchi2
stefano.bistarelli@unipg.it francesco.santini@unipg.it carlo.taticchi@gssi.it
Thanks for your attention!
Questions?
1Universit`a degli Studi di Perugia, Italy
2Gran Sasso Science Institute (GSSI), L’Aquila, Italy

More Related Content

PDF
A Concurrent Language for Argumentation
PDF
Introducing a Tool for Concurrent Argumentation
PDF
A Concurrent Language for Argumentation: Preliminary Notes
PDF
Looking for Invariant Operators in Argumentation
PDF
Timed Concurrent Language for Argumentation
PDF
Timed Concurrent Language for Argumentation: an Interleaving Approach
PDF
A Unifying Four-State Labelling Semantics for Bridging Abstract Argumentation...
PDF
A Concurrent Argumentation Language for Negotiation and Debating
A Concurrent Language for Argumentation
Introducing a Tool for Concurrent Argumentation
A Concurrent Language for Argumentation: Preliminary Notes
Looking for Invariant Operators in Argumentation
Timed Concurrent Language for Argumentation
Timed Concurrent Language for Argumentation: an Interleaving Approach
A Unifying Four-State Labelling Semantics for Bridging Abstract Argumentation...
A Concurrent Argumentation Language for Negotiation and Debating

What's hot (20)

PDF
Strategic Argumentation is NP-complete
PPT
A Study Of Statistical Models For Query Translation :Finding A Good Unit Of T...
 
PPT
Mapping FAO’s AGROVOC Thesaurus
PPTX
Csr2011 june17 15_15_kaminski
PDF
ThesisDefenseRR
PPT
Algorithm
PPT
Logic Programming and Prolog
PPT
Sdd Syntax Descriptions
PDF
Complex Arguments in Adpositional Argumentation
PDF
Minimizing DFA
PPTX
Deterministic context free grammars &non-deterministic
PDF
Slides Workshopon Explainable Logic-Based Knowledge Representation (XLoKR 2020)
PDF
CodeJugalbandi-Expression-Problem-HealthyCode-Magazine#Jan-2015-Issue
PDF
Fundamentals of Computing and C Programming - Part 2
DOCX
Mit203 analysis and design of algorithms
PDF
Semantic Analysis of a C Program
PDF
Theory of Computation Lecture Notes
PDF
Fundamentals of Computing and C Programming - Part 1
PDF
ANOTHER PROOF OF THE DENUMERABILITY OF THE COMPLEX NUMBERS
PDF
Strategic Argumentation is NP-complete
A Study Of Statistical Models For Query Translation :Finding A Good Unit Of T...
 
Mapping FAO’s AGROVOC Thesaurus
Csr2011 june17 15_15_kaminski
ThesisDefenseRR
Algorithm
Logic Programming and Prolog
Sdd Syntax Descriptions
Complex Arguments in Adpositional Argumentation
Minimizing DFA
Deterministic context free grammars &non-deterministic
Slides Workshopon Explainable Logic-Based Knowledge Representation (XLoKR 2020)
CodeJugalbandi-Expression-Problem-HealthyCode-Magazine#Jan-2015-Issue
Fundamentals of Computing and C Programming - Part 2
Mit203 analysis and design of algorithms
Semantic Analysis of a C Program
Theory of Computation Lecture Notes
Fundamentals of Computing and C Programming - Part 1
ANOTHER PROOF OF THE DENUMERABILITY OF THE COMPLEX NUMBERS
Ad

Similar to Looking for Invariant Operators in Argumentation (11)

PDF
Handling Dynamic Aspects of Argumentation
PDF
AN IMPLEMENTATION, EMPIRICAL EVALUATION AND PROPOSED IMPROVEMENT FOR BIDIRECT...
PDF
A Matrix Based Approach for Weighted Argumentation Frameworks
PPTX
AI_session 22 inference and unification.pptx
PPTX
22PCOAM11 Unit 3: Session 14 Inference and Unification.pptx
PDF
CS6660-COMPILER DESIGN-1368874055-CD NOTES-55-150.pdf
PDF
AI Lesson 13
PDF
Argumentation in Artificial Intelligence
PDF
Link Discovery Tutorial Part II: Accuracy
PPTX
Syntax Analysis in Compiler Design
PDF
Rule-based reasoning in the Semantic Web
Handling Dynamic Aspects of Argumentation
AN IMPLEMENTATION, EMPIRICAL EVALUATION AND PROPOSED IMPROVEMENT FOR BIDIRECT...
A Matrix Based Approach for Weighted Argumentation Frameworks
AI_session 22 inference and unification.pptx
22PCOAM11 Unit 3: Session 14 Inference and Unification.pptx
CS6660-COMPILER DESIGN-1368874055-CD NOTES-55-150.pdf
AI Lesson 13
Argumentation in Artificial Intelligence
Link Discovery Tutorial Part II: Accuracy
Syntax Analysis in Compiler Design
Rule-based reasoning in the Semantic Web
Ad

More from Carlo Taticchi (20)

PDF
Empowering Public Interest Communication with Argumentation - Project Overview
PDF
Modelling Dialogues in a Concurrent Language for Argumentation
PDF
Preserving Privacy in a (Timed) Concurrent Language for Argumentation
PDF
A Semantics-Aware Evaluation Order for Abstract Argumentation Frameworks
PDF
On the Role of Local Arguments in the (Timed) Concurrent Language for Argumen...
PDF
Session3_ 52_Taticchi.pdf
PDF
Arg-XAI: a Tool for Explaining Machine Learning Results
PDF
A Four-State Labelling Semantics for Weighted Argumentation Frameworks
PDF
Extending Labelling Semantics to Weighted Argumentation Frameworks
PDF
A Chatbot Extended with Argumentation
PDF
Concurrent Argumentation with Time: an Overview
PDF
Ranking-Based Semantics from the Perspective of Claims
PDF
A Labelling Semantics for Weighted Argumentation Frameworks
PDF
Implementing Ranking-Based Semantics in ConArg
PDF
Third International Competition on Computational Models of Argumentation
PDF
Containerisation and Dynamic Frameworks in ICCMA’19
PDF
Preliminary Study on Reinstatement Labelling for Weighted Argumentation Frame...
PDF
A Tool For Ranking Arguments Through Voting-Games Power Indexes
PDF
Acceptability Paradigms in Abstract Argumentation Frameworks
PDF
A Cooperative-game Approach to Share Acceptability and Rank Arguments
Empowering Public Interest Communication with Argumentation - Project Overview
Modelling Dialogues in a Concurrent Language for Argumentation
Preserving Privacy in a (Timed) Concurrent Language for Argumentation
A Semantics-Aware Evaluation Order for Abstract Argumentation Frameworks
On the Role of Local Arguments in the (Timed) Concurrent Language for Argumen...
Session3_ 52_Taticchi.pdf
Arg-XAI: a Tool for Explaining Machine Learning Results
A Four-State Labelling Semantics for Weighted Argumentation Frameworks
Extending Labelling Semantics to Weighted Argumentation Frameworks
A Chatbot Extended with Argumentation
Concurrent Argumentation with Time: an Overview
Ranking-Based Semantics from the Perspective of Claims
A Labelling Semantics for Weighted Argumentation Frameworks
Implementing Ranking-Based Semantics in ConArg
Third International Competition on Computational Models of Argumentation
Containerisation and Dynamic Frameworks in ICCMA’19
Preliminary Study on Reinstatement Labelling for Weighted Argumentation Frame...
A Tool For Ranking Arguments Through Voting-Games Power Indexes
Acceptability Paradigms in Abstract Argumentation Frameworks
A Cooperative-game Approach to Share Acceptability and Rank Arguments

Recently uploaded (20)

PPTX
ANEMIA WITH LEUKOPENIA MDS 07_25.pptx htggtftgt fredrctvg
PPT
The World of Physical Science, • Labs: Safety Simulation, Measurement Practice
PDF
HPLC-PPT.docx high performance liquid chromatography
PPTX
Classification Systems_TAXONOMY_SCIENCE8.pptx
PDF
IFIT3 RNA-binding activity primores influenza A viruz infection and translati...
PPTX
Protein & Amino Acid Structures Levels of protein structure (primary, seconda...
PPTX
Cell Membrane: Structure, Composition & Functions
PPTX
G5Q1W8 PPT SCIENCE.pptx 2025-2026 GRADE 5
PPTX
ognitive-behavioral therapy, mindfulness-based approaches, coping skills trai...
PPTX
INTRODUCTION TO EVS | Concept of sustainability
PPTX
GEN. BIO 1 - CELL TYPES & CELL MODIFICATIONS
PPTX
ECG_Course_Presentation د.محمد صقران ppt
PPT
protein biochemistry.ppt for university classes
PPT
POSITIONING IN OPERATION THEATRE ROOM.ppt
PPTX
microscope-Lecturecjchchchchcuvuvhc.pptx
PPTX
cpcsea ppt.pptxssssssssssssssjjdjdndndddd
PPTX
BIOMOLECULES PPT........................
PDF
SEHH2274 Organic Chemistry Notes 1 Structure and Bonding.pdf
PDF
An interstellar mission to test astrophysical black holes
PPTX
DRUG THERAPY FOR SHOCK gjjjgfhhhhh.pptx.
ANEMIA WITH LEUKOPENIA MDS 07_25.pptx htggtftgt fredrctvg
The World of Physical Science, • Labs: Safety Simulation, Measurement Practice
HPLC-PPT.docx high performance liquid chromatography
Classification Systems_TAXONOMY_SCIENCE8.pptx
IFIT3 RNA-binding activity primores influenza A viruz infection and translati...
Protein & Amino Acid Structures Levels of protein structure (primary, seconda...
Cell Membrane: Structure, Composition & Functions
G5Q1W8 PPT SCIENCE.pptx 2025-2026 GRADE 5
ognitive-behavioral therapy, mindfulness-based approaches, coping skills trai...
INTRODUCTION TO EVS | Concept of sustainability
GEN. BIO 1 - CELL TYPES & CELL MODIFICATIONS
ECG_Course_Presentation د.محمد صقران ppt
protein biochemistry.ppt for university classes
POSITIONING IN OPERATION THEATRE ROOM.ppt
microscope-Lecturecjchchchchcuvuvhc.pptx
cpcsea ppt.pptxssssssssssssssjjdjdndndddd
BIOMOLECULES PPT........................
SEHH2274 Organic Chemistry Notes 1 Structure and Bonding.pdf
An interstellar mission to test astrophysical black holes
DRUG THERAPY FOR SHOCK gjjjgfhhhhh.pptx.

Looking for Invariant Operators in Argumentation

  • 1. Looking for Invariant Operators in Argumentation Stefano Bistarelli1 Francesco Santini1 Carlo Taticchi2 stefano.bistarelli@unipg.it francesco.santini@unipg.it carlo.taticchi@gssi.it 1Universit`a degli Studi di Perugia, Italy 2Gran Sasso Science Institute (GSSI), L’Aquila, Italy
  • 6. Index 1 Background - Argumentation Frameworks 2 Definition - Invariant Operators 3 Conclusion and Future Work Carlo Taticchi Looking for Invariant Operators in Argumentation 3 / 14
  • 7. Argumentation Frameworks1 A human-like fashion to deal with knowledge Definition (AF) An Abstract Argumentation Framework is a pair G = A, R where A is a set of arguments and R is a binary relation on A. 1 PHAN MINH DUNG. On the Acceptability of Arguments and its Fundamental Role in Nonmonotonic Reasoning, Logic Programming and n-Person Games. Artif. Intell., 77(2):321–358, 1995. Carlo Taticchi Looking for Invariant Operators in Argumentation 4 / 14
  • 8. Sets of Extensions Definition (Conflict-free extensions) Let G = A, R be an AF. A set E ⊆ A is conflict-free in G if there are no a, b ∈ A | (a, b) ∈ R. Carlo Taticchi Looking for Invariant Operators in Argumentation 5 / 14
  • 9. Sets of Extensions Definition (Conflict-free extensions) Let G = A, R be an AF. A set E ⊆ A is conflict-free in G if there are no a, b ∈ A | (a, b) ∈ R. Scf (G) = {{}, {1}, {2}, {3}, {4}, {5}, {1, 2}, {1,4}, {1, 5}, {2,5}, {3, 5}, {1, 2, 5}} Carlo Taticchi Looking for Invariant Operators in Argumentation 5 / 14
  • 10. Dung’s Semantics Definition (Admissible Semantics) Let G = A, R be an AF. A set E ⊆ A is admissible in G if E is conflict-free and each a ∈ E is defended by E. Sadm(G) = {{}, {1},{2}, {3}, {4}, {5}, {1, 2},{1,4},{1, 5}, {2, 5}, {3, 5}, {1, 2, 5}} Carlo Taticchi Looking for Invariant Operators in Argumentation 6 / 14
  • 11. Dung’s Semantics Definition (Admissible Semantics) Let G = A, R be an AF. A set E ⊆ A is admissible in G if E is conflict-free and each a ∈ E is defended by E. Similar definitions for complete, grounded, preferred and stable. Carlo Taticchi Looking for Invariant Operators in Argumentation 6 / 14
  • 12. Reinstatement Labelling2 Every argument can be labelled in, out or undec. Example (Labelling of an AF) 2 MARTIN CAMINADA. On the Issue of Reinstatement in Argumentation. Logics in Artificial Intelligence: 10th European Conference, JELIA 2006 Liverpool, UK, September 13-15, 2006 Proceedings, pages 111–123, 2006. Carlo Taticchi Looking for Invariant Operators in Argumentation 7 / 14
  • 13. Robustness3,4 A property of an AF to withstand changes Tries to answer the following questions: • Is it possible to change the outcome of a debate according to a particular semantics or meaning? • If so, how easy could it be to perform such change? • And which consequences does it bring? 3 S. BISTARELLI, F. FALOCI, F. SANTINI, AND C. TATICCHI. Robustness in abstract argumentation frameworks. In Proceedings of the 29th International Florida Artificial Intelligence Research Society Conf. FLAIRS, page 703, 2016. 4 CARLO TATICCHI. A Study of Robustness in Abstract Argumentation Frameworks. Proceedings of the Doctoral Consortium of AI*IA 2016, Genova, Italy, November 29, 2016, pages 11–16. CEUR-WS.org, 2016. Carlo Taticchi Looking for Invariant Operators in Argumentation 8 / 14
  • 14. Modifying Operators Allow to construct AFs adding an attack relations at time. Definition (Modifying operator) Let G = A, R ∈ Fn be an AF. A modifying operator is a function m : Fn → Fn such that m(G) = A, m(R) , where m(R) ⊇ R. Carlo Taticchi Looking for Invariant Operators in Argumentation 9 / 14
  • 15. Modifying Operators Allow to construct AFs adding an attack relations at time. Definition (Modifying operator) Let G = A, R ∈ Fn be an AF. A modifying operator is a function m : Fn → Fn such that m(G) = A, m(R) , where m(R) ⊇ R. Carlo Taticchi Looking for Invariant Operators in Argumentation 9 / 14
  • 16. Modifying Operators Allow to construct AFs adding an attack relations at time. Definition (Modifying operator) Let G = A, R ∈ Fn be an AF. A modifying operator is a function m : Fn → Fn such that m(G) = A, m(R) , where m(R) ⊇ R. Carlo Taticchi Looking for Invariant Operators in Argumentation 9 / 14
  • 17. Comparing AFs We have defined a partial order on the set Fn of all AFs with n arguments and all possible combinations of attack relations. Definition (AFs inclusion w.r.t. attacks) Let G1 = A, R1 and G2 = A, R2 be two AFs. We say that G1 is included in G2 w.r.t. attacks if R1 ⊆ R2 and we write G1 ≤A G2. Figure: G1 Figure: G2 Carlo Taticchi Looking for Invariant Operators in Argumentation 10 / 14
  • 18. Comparing Semantics Definition (Semantics inclusion) Let S and S be two sets of extensions. We say that S ⊆ S if and only if ∀E ∈ S ∃E ∈ S | E ⊆ E . Figure: Sadm(G1) = {{}, {1}, {1, 3}} Figure: Sadm(G2) = {{}, {1}} Carlo Taticchi Looking for Invariant Operators in Argumentation 11 / 14
  • 19. Invariant Operators for Admissible Semantics Example (AF with 4 arguments) Figure: Sadm = {{}, {1}, {4}, {1, 4}} Carlo Taticchi Looking for Invariant Operators in Argumentation 12 / 14
  • 20. Invariant Operators for Admissible Semantics Example (AF with 4 arguments) Figure: Sadm = {{}, {1}, {4}, {1, 4}} 2 → 1: ∃c = b such that there is an odd length sequence of attacks from c to a, but not from a to c. Carlo Taticchi Looking for Invariant Operators in Argumentation 12 / 14
  • 21. Invariant Operators for Admissible Semantics Example (AF with 4 arguments) Figure: Sadm = {{}, {1}, {4}, {1, 4}} 2 → 3: there is no odd length sequence of attacks from b to a. Carlo Taticchi Looking for Invariant Operators in Argumentation 12 / 14
  • 22. Invariant Operators for Admissible Semantics Example (AF with 4 arguments) Figure: Sadm = {{}, {1}, {4}, {1, 4}} 4 → 3: c ∈ in(L) such that (b, c) ∈ R. Carlo Taticchi Looking for Invariant Operators in Argumentation 12 / 14
  • 23. Invariant Operators for Admissible Semantics Example (AF with 4 arguments) Figure: Sadm = {{}, {1}, {4}, {1, 4}} Carlo Taticchi Looking for Invariant Operators in Argumentation 12 / 14
  • 24. Summing Up Invariant local expansion operators for the admissible semantics allow for: • adding an attack in an AF, while • maintaining the admissible semantics unchanged a → b Condition in → in never possible in → out c ∈ in(L) | (b, c) ∈ R in → undec c ∈ undec(L) | (c, c) /∈ R and (b, c) ∈ R out → in (b, a) ∈ R or ∃c ∈ out | (c, b) ∈ R out → in there is no odd length sequence of attacks from b to a out → out or ∃c = b | there is an odd length sequence of attacks undec from c to a, but not from a to c undec → in never possible undec → out no condition required undec → undec no condition required Carlo Taticchi Looking for Invariant Operators in Argumentation 13 / 14
  • 25. Further Research Next possible steps: • To design invariant operators with respect to the complete, grounded, preferred, semi-stable, and stable semantics. • To find the essential extensions for which every change inside modifies the semantics, while changes outside do not. • To study local expansion operators also for semiring-based weighted AFs5. • To consider different notions of equivalence6 (e.g. local equivalence) and additional modifications of AFs (as the deletion of attack or the addition/removal of arguments). 5 S. BISTARELLI, F. ROSSI, AND F. SANTINI. A Collective Defence Against Grouped Attacks for Weighted Abstract Argumentation Frameworks. FLAIRS Conference 2016: 638-643 6 E. OIKARINEN, AND S. WOLTRAN. Characterizing strong equivalence for argumentation frameworks. 2011. Artificial Intelligence 175(14-15):1985–2009. Carlo Taticchi Looking for Invariant Operators in Argumentation 14 / 14
  • 26. Further Research Next possible steps: • To design invariant operators with respect to the complete, grounded, preferred, semi-stable, and stable semantics. • To find the essential extensions for which every change inside modifies the semantics, while changes outside do not. • To study local expansion operators also for semiring-based weighted AFs5. • To consider different notions of equivalence6 (e.g. local equivalence) and additional modifications of AFs (as the deletion of attack or the addition/removal of arguments). 5 S. BISTARELLI, F. ROSSI, AND F. SANTINI. A Collective Defence Against Grouped Attacks for Weighted Abstract Argumentation Frameworks. FLAIRS Conference 2016: 638-643 6 E. OIKARINEN, AND S. WOLTRAN. Characterizing strong equivalence for argumentation frameworks. 2011. Artificial Intelligence 175(14-15):1985–2009. Carlo Taticchi Looking for Invariant Operators in Argumentation 14 / 14
  • 27. Further Research Next possible steps: • To design invariant operators with respect to the complete, grounded, preferred, semi-stable, and stable semantics. • To find the essential extensions for which every change inside modifies the semantics, while changes outside do not. • To study local expansion operators also for semiring-based weighted AFs5. • To consider different notions of equivalence6 (e.g. local equivalence) and additional modifications of AFs (as the deletion of attack or the addition/removal of arguments). 5 S. BISTARELLI, F. ROSSI, AND F. SANTINI. A Collective Defence Against Grouped Attacks for Weighted Abstract Argumentation Frameworks. FLAIRS Conference 2016: 638-643 6 E. OIKARINEN, AND S. WOLTRAN. Characterizing strong equivalence for argumentation frameworks. 2011. Artificial Intelligence 175(14-15):1985–2009. Carlo Taticchi Looking for Invariant Operators in Argumentation 14 / 14
  • 28. Further Research Next possible steps: • To design invariant operators with respect to the complete, grounded, preferred, semi-stable, and stable semantics. • To find the essential extensions for which every change inside modifies the semantics, while changes outside do not. • To study local expansion operators also for semiring-based weighted AFs5. • To consider different notions of equivalence6 (e.g. local equivalence) and additional modifications of AFs (as the deletion of attack or the addition/removal of arguments). 5 S. BISTARELLI, F. ROSSI, AND F. SANTINI. A Collective Defence Against Grouped Attacks for Weighted Abstract Argumentation Frameworks. FLAIRS Conference 2016: 638-643 6 E. OIKARINEN, AND S. WOLTRAN. Characterizing strong equivalence for argumentation frameworks. 2011. Artificial Intelligence 175(14-15):1985–2009. Carlo Taticchi Looking for Invariant Operators in Argumentation 14 / 14
  • 29. Looking for Invariant Operators in Argumentation Stefano Bistarelli1 Francesco Santini1 Carlo Taticchi2 stefano.bistarelli@unipg.it francesco.santini@unipg.it carlo.taticchi@gssi.it Thanks for your attention! Questions? 1Universit`a degli Studi di Perugia, Italy 2Gran Sasso Science Institute (GSSI), L’Aquila, Italy