Modelling Dialogues in a Concurrent Language for Argumentation

1. Modelling Dialogues in a Concurrent Language for Argumentation Stefano Bistarelli, Maria Chiara Meo, Carlo Taticchi LPNMR 2024

2. LPNMR 2024 Modelling Dialogues in a Concurrent Language for Argumentation Some motivation Paul: My car is safe. Olga: Why is your car safe? Paul: Since it has an airbag. Olga: That is true, but this does not make your car safe. Paul: Why does that not make my care safe? Olga: Newspapers recently reported on airbags exploding without cause. Paul: That does not prove anything since newspaper reports are unreliable. Olga: Still, your car is not safe since its maximum speed is very high. Paul: OK, I was wrong: my car is not safe. 2 Prakken, H.: Formal systems for persuasion dialogue. Knowl. Eng. Rev. 21(2), 163–188 (2006)

3. LPNMR 2024 Modelling Dialogues in a Concurrent Language for Argumentation Some motivation Paul: My car is safe. Olga: Why is your car safe? Paul: Since it has an airbag. Olga: That is true, but this does not make your car safe. Paul: Why does that not make my care safe? Olga: Newspapers recently reported on airbags exploding without cause. Paul: That does not prove anything since newspaper reports are unreliable. Olga: Still, your car is not safe since its maximum speed is very high. Paul: OK, I was wrong: my car is not safe. 2 Prakken, H.: Formal systems for persuasion dialogue. Knowl. Eng. Rev. 21(2), 163–188 (2006)

4. LPNMR 2024 Modelling Dialogues in a Concurrent Language for Argumentation Formal operational models cannot be directly used to represent debates and reason with their outcomes 3

5. LPNMR 2024 Modelling Dialogues in a Concurrent Language for Argumentation Goals • Providing domain experts with a tool for representing and automating reasoning processes • Reproducing the complex aspects of human communication • Conducting dialogues between several agents • Moving beyond simple turn-taking to more intricate, concurrent communication systems 4

6. Overview • Motivations • Speech Acts & Abstract Argumentation • Dialogic Concurrent Language for Argumentation • Operational semantics • Conclusion

7. LPNMR 2024 Modelling Dialogues in a Concurrent Language for Argumentation Speech Acts for Communication Languages • Linguistic expression that conveys information • Some examples: • claim(φ): claiming that a formula φ holds 6 “The new environmental policy will significantly reduce pollution”

8. LPNMR 2024 Modelling Dialogues in a Concurrent Language for Argumentation Speech Acts for Communication Languages • Linguistic expression that conveys information • Some examples: • claim(φ): claiming that a formula φ holds • counter(φ, ψ): stating φ as a counterclaim to ψ 6 “The new environmental policy will significantly reduce pollution” “While the policy aims to reduce pollution, it overlooks the increase in manufacturing waste it could cause”

9. LPNMR 2024 Modelling Dialogues in a Concurrent Language for Argumentation Speech Acts for Communication Languages • Linguistic expression that conveys information • Some examples: • claim(φ): claiming that a formula φ holds • counter(φ, ψ): stating φ as a counterclaim to ψ • why(φ): asking for a claim supporting φ 6 “The new environmental policy will significantly reduce pollution” “While the policy aims to reduce pollution, it overlooks the increase in manufacturing waste it could cause” “What makes you believe that the new environmental policy will be effective?”

10. LPNMR 2024 Modelling Dialogues in a Concurrent Language for Argumentation Speech Acts for Communication Languages • Linguistic expression that conveys information • Some examples: • claim(φ): claiming that a formula φ holds • counter(φ, ψ): stating φ as a counterclaim to ψ • why(φ): asking for a claim supporting φ • argue(φ, ψ): providing a claim φ supporting ψ 6 “The new environmental policy will significantly reduce pollution” “While the policy aims to reduce pollution, it overlooks the increase in manufacturing waste it could cause” “Research from multiple cities demonstrates that similar policies have decreased urban pollution by up to 40%” “What makes you believe that the new environmental policy will be effective?”

11. LPNMR 2024 Modelling Dialogues in a Concurrent Language for Argumentation Speech Acts for Communication Languages • Linguistic expression that conveys information • Some examples: • claim(φ): claiming that a formula φ holds • counter(φ, ψ): stating φ as a counterclaim to ψ • why(φ): asking for a claim supporting φ • argue(φ, ψ): providing a claim φ supporting ψ • concede(φ): accepting that φ holds 6 “The new environmental policy will significantly reduce pollution” “While the policy aims to reduce pollution, it overlooks the increase in manufacturing waste it could cause” “Research from multiple cities demonstrates that similar policies have decreased urban pollution by up to 40%” “I recognise your point about production waste; we need to address this unintended consequence” “What makes you believe that the new environmental policy will be effective?”

12. LPNMR 2024 Modelling Dialogues in a Concurrent Language for Argumentation Speech Acts for Communication Languages • Linguistic expression that conveys information • Some examples: • claim(φ): claiming that a formula φ holds • counter(φ, ψ): stating φ as a counterclaim to ψ • why(φ): asking for a claim supporting φ • argue(φ, ψ): providing a claim φ supporting ψ • concede(φ): accepting that φ holds • retract(φ): retracting φ, which had previously been claimed 6 “The new environmental policy will significantly reduce pollution” “While the policy aims to reduce pollution, it overlooks the increase in manufacturing waste it could cause” “Research from multiple cities demonstrates that similar policies have decreased urban pollution by up to 40%” “I recognise your point about production waste; we need to address this unintended consequence” “Considering the recent data, I rescind my initial claim regarding the success of the policy” “What makes you believe that the new environmental policy will be effective?”

13. LPNMR 2024 Modelling Dialogues in a Concurrent Language for Argumentation • AF = ⟨Arg, Att⟩ • Compute acceptability (labelling semantics) • • L(a) = in ⟹ ∀b ∈ Arg ∣ (b, a) ∈ Att . L(b) = out L(a) = out ⟺ ∃b ∈ Arg . (b, a) ∈ Att ∧ L(b) = in Abstract Argumentation Frameworks 7 Baroni, P., Caminada, M., Giacomin, M.: An introduction to argumentation semantics. Knowl. Eng. Rev. 26(4), 365–410 (2011)

14. LPNMR 2024 Modelling Dialogues in a Concurrent Language for Argumentation • AF = ⟨Arg, Att⟩ • Compute acceptability (labelling semantics) • • L(a) = in ⟹ ∀b ∈ Arg ∣ (b, a) ∈ Att . L(b) = out L(a) = out ⟺ ∃b ∈ Arg . (b, a) ∈ Att ∧ L(b) = in Abstract Argumentation Frameworks 7 Baroni, P., Caminada, M., Giacomin, M.: An introduction to argumentation semantics. Knowl. Eng. Rev. 26(4), 365–410 (2011)

15. LPNMR 2024 Modelling Dialogues in a Concurrent Language for Argumentation • AF = ⟨Arg, Att⟩ • Compute acceptability (labelling semantics) • • L(a) = in ⟹ ∀b ∈ Arg ∣ (b, a) ∈ Att . L(b) = out L(a) = out ⟺ ∃b ∈ Arg . (b, a) ∈ Att ∧ L(b) = in Abstract Argumentation Frameworks 7 • BAF = ⟨Arg, Att, Supp⟩ Baroni, P., Caminada, M., Giacomin, M.: An introduction to argumentation semantics. Knowl. Eng. Rev. 26(4), 365–410 (2011)

16. LPNMR 2024 Modelling Dialogues in a Concurrent Language for Argumentation Dialogic Concurrent Language for Argumentation (DICLA) 8 D ::= claim(a) ∣ counter(a, b) ∣ why(a) ∣ argue(a, b) ∣ concede(a) ∣ retract(a) ∣ D∥D ∣ D → D •Paul: My car is safe. •Olga: Why is your car safe? •Paul: Since it has an airbag. •Olga: That is true, but this does not make your car safe. •Paul: Why does that not make my care safe? •Olga: Newspapers recently reported on airbags exploding without cause. •Paul: That does not prove anything since newspaper reports are unreliable. •Olga: Still, your car is not safe since its maximum speed is very high. •Paul: OK, I was wrong: my car is not safe.

25. LPNMR 2024 Modelling Dialogues in a Concurrent Language for Argumentation Low level implementation • Timed Concurrent Language for Argumentation (TCLA) • Parallel composition obtained through interleaving • Time progresses uniformly for all processes within the context • Guarded actions can suspend the execution A∥A ∥ 9 Bistarelli S., Taticchi C., Meo, M. C.: An Interleaving Semantics of the Timed Concurrent Language for Argumentation to Model Debates and Dialogue Games. Theory Pract. Log. Program. 23(6): 1307-1333 (2023)

26. LPNMR 2024 Modelling Dialogues in a Concurrent Language for Argumentation DICLA operational semantics Translation of the speech acts into TCLA constructs 10 D ::= claim(a) ∣ counter(a, b) ∣ why(a) ∣ argue(a, b) ∣ concede(a) ∣ retract(a) ∣ D∥D ∣ D → D

28. LPNMR 2024 Modelling Dialogues in a Concurrent Language for Argumentation DICLA operational semantics Translation of the speech acts into TCLA constructs 10 stands for ∃x ∈ S ∣ p(x) p(a1) + … + p(an) D ::= claim(a) ∣ counter(a, b) ∣ why(a) ∣ argue(a, b) ∣ concede(a) ∣ retract(a) ∣ D∥D ∣ D → D

30. LPNMR 2024 Modelling Dialogues in a Concurrent Language for Argumentation DICLA operational semantics Translation of the speech acts into TCLA constructs 10 stands for ∀x ∈ S ∣ p(x) p(a1) ∥ … ∥ p(an) D ::= claim(a) ∣ counter(a, b) ∣ why(a) ∣ argue(a, b) ∣ concede(a) ∣ retract(a) ∣ D∥D ∣ D → D

32. LPNMR 2024 Modelling Dialogues in a Concurrent Language for Argumentation DICLA parallel operator • Exploits TLCA interleaving • Only one action can be executed at a time • All other actions in the || context will have their timeout decreased 11 D ::= claim(a) ∣ counter(a, b) ∣ why(a) ∣ argue(a, b) ∣ concede(a) ∣ retract(a) ∣ D∥D ∣ D → D

33. LPNMR 2024 Modelling Dialogues in a Concurrent Language for Argumentation DICLA sequential operator • Translated into TCLA constructs • In the program , agent will always be executed after , and likewise after D1 → D2 ∥ D3 → D4 D2 D1 D4 D3 12 D ::= claim(a) ∣ counter(a, b) ∣ why(a) ∣ argue(a, b) ∣ concede(a) ∣ retract(a) ∣ D∥D ∣ D → D

34. LPNMR 2024 Modelling Dialogues in a Concurrent Language for Argumentation Conclusion and future work • DICLA integrates concurrent programming paradigms with computational argumentation • Implements high-level propositions to model Speech Acts • Operational semantics based on TCLA 13

35. LPNMR 2024 Modelling Dialogues in a Concurrent Language for Argumentation Conclusion and future work • DICLA integrates concurrent programming paradigms with computational argumentation • Implements high-level propositions to model Speech Acts • Operational semantics based on TCLA • Provide agents with a local store to enrich the behaviour of “why(a)” and “concede(a)” • Enforce protocol constraints — prevent retraction of statements made by others • Time-based propositions to be handled within a given time frame — why7(a) 13

36. Modelling Dialogues in a Concurrent Language for Argumentation Stefano Bistarelli, Maria Chiara Meo, Carlo Taticchi Thank you for your attention!

Modelling Dialogues in a Concurrent Language for Argumentation

More Related Content

More from Carlo Taticchi (20)

Recently uploaded (20)

Modelling Dialogues in a Concurrent Language for Argumentation