On the Role of Local Arguments in the (Timed) Concurrent Language for Argumentation
0 likes8 views
This document presents the Timed Concurrent Language for Argumentation (TCLA), which allows modeling of agent interactions and strategic argumentation. TCLA introduces local arguments that agents can use to reason privately with hidden information. It also features timeout environments and true concurrency semantics to model parallel operations. Examples demonstrate how TCLA can be used for strategic persuasion and decision making while preserving privacy. Future work is discussed to further develop the interleaving semantics and modeling of agent ownership and strategies.
On the Role of Local Arguments in the (Timed) Concurrent Language for Argumentation
- 1. On the Role of Local Arguments in the (Timed) Concurrent Language for Argumentation Stefano Bistarelli, Maria Chiara Meo and Carlo Taticchi
- 2. AI^3@AIxIA 2023 Deriving Dependency Graphs from Abstract Argumentation Frameworks Overview • Timed Concurrent Language for Argumentation ‣ Syntax ‣ Operational Semantics • Timeout Environments • Locality • Examples 2
- 3. AI^3@AIxIA 2023 Deriving Dependency Graphs from Abstract Argumentation Frameworks tcla Syntax 3
- 4. AI^3@AIxIA 2023 Deriving Dependency Graphs from Abstract Argumentation Frameworks tcla Syntax 3
- 5. AI^3@AIxIA 2023 Deriving Dependency Graphs from Abstract Argumentation Frameworks tcla Syntax 3
- 6. AI^3@AIxIA 2023 Deriving Dependency Graphs from Abstract Argumentation Frameworks tcla Syntax 3
- 7. AI^3@AIxIA 2023 Deriving Dependency Graphs from Abstract Argumentation Frameworks tcla Syntax 3
- 8. AI^3@AIxIA 2023 Deriving Dependency Graphs from Abstract Argumentation Frameworks tcla Syntax 3
- 9. AI^3@AIxIA 2023 Deriving Dependency Graphs from Abstract Argumentation Frameworks tcla Syntax 3
- 10. AI^3@AIxIA 2023 Deriving Dependency Graphs from Abstract Argumentation Frameworks Handling Parallelism • We used true concurrency ‣ Infinite number of processors ‣ All concurrent agents executed simultaneously 4 CPU CPU CPU
- 11. AI^3@AIxIA 2023 Deriving Dependency Graphs from Abstract Argumentation Frameworks Handling Parallelism • We used true concurrency ‣ Infinite number of processors ‣ All concurrent agents executed simultaneously • Alternative: interleaving (work in progress) ‣ 1 processor ‣ 1 operation at a time 4 CPU CPU CPU CPU
- 12. AI^3@AIxIA 2023 Deriving Dependency Graphs from Abstract Argumentation Frameworks Timeout Environments • The passing of time between agents is handled on the same global clock via a timeout environment 𝑇 : I ⇀ N ∪ {∞} 5
- 13. AI^3@AIxIA 2023 Deriving Dependency Graphs from Abstract Argumentation Frameworks Timeout Environments • The passing of time between agents is handled on the same global clock via a timeout environment 𝑇 : I ⇀ N ∪ {∞} 5 I 𝑇(I) Agent A 4 Agent B 1 Agent C 2
- 14. AI^3@AIxIA 2023 Deriving Dependency Graphs from Abstract Argumentation Frameworks Timeout Environments • The passing of time between agents is handled on the same global clock via a timeout environment 𝑇 : I ⇀ N ∪ {∞} • We can update/decrement 𝑇 5 I 𝑇(I) Agent A 4 Agent B 1 Agent C 2
- 15. AI^3@AIxIA 2023 Deriving Dependency Graphs from Abstract Argumentation Frameworks True Concurrency Semantics • *(𝐹, 𝐹′, 𝐹′′) := (𝐹′ ∩ 𝐹′′) ∪ ((𝐹′ ∪ 𝐹′′ ) ∖ 𝐹) • Example: add({𝑎}, {}) → 𝑠𝑢𝑐𝑐𝑒𝑠𝑠 ‖ rmv({𝑎}, {}) → 𝑠𝑢𝑐𝑐𝑒𝑠𝑠 6
- 16. AI^3@AIxIA 2023 Deriving Dependency Graphs from Abstract Argumentation Frameworks Check Semantics 7
- 17. AI^3@AIxIA 2023 Deriving Dependency Graphs from Abstract Argumentation Frameworks Check Semantics 7
- 18. AI^3@AIxIA 2023 Deriving Dependency Graphs from Abstract Argumentation Frameworks Check Semantics 7
- 19. AI^3@AIxIA 2023 Deriving Dependency Graphs from Abstract Argumentation Frameworks Check Semantics 7
- 20. AI^3@AIxIA 2023 Deriving Dependency Graphs from Abstract Argumentation Frameworks Check Semantics 7 Test semantics are similar to the check one but for the conditions to satisfy
- 21. AI^3@AIxIA 2023 Deriving Dependency Graphs from Abstract Argumentation Frameworks Nondeterminism and If-Then-Else Semantics 8
- 22. AI^3@AIxIA 2023 Deriving Dependency Graphs from Abstract Argumentation Frameworks Nondeterminism and If-Then-Else Semantics 8
- 23. AI^3@AIxIA 2023 Deriving Dependency Graphs from Abstract Argumentation Frameworks Locality Semantics • new 𝑆 in 𝐴 behaves like 𝐴 where arguments in 𝑆 are local to 𝐴 • 𝐴𝐹𝑙𝑜𝑐 contains information on 𝑆 which is hidden from the external 𝐴𝐹 9
- 24. AI^3@AIxIA 2023 Deriving Dependency Graphs from Abstract Argumentation Frameworks Locality Semantics 9 𝐴𝐹 ↑ 𝑆 𝑆
- 25. AI^3@AIxIA 2023 Deriving Dependency Graphs from Abstract Argumentation Frameworks • Strategic Argumentation: disclose those right arguments capable of persuading the other player Example 1 10
- 26. AI^3@AIxIA 2023 Deriving Dependency Graphs from Abstract Argumentation Frameworks • Strategic Argumentation: disclose those right arguments capable of persuading the other player Example 1 10
- 27. AI^3@AIxIA 2023 Deriving Dependency Graphs from Abstract Argumentation Frameworks • Strategic Argumentation: disclose those right arguments capable of persuading the other player Example 1 10
- 28. AI^3@AIxIA 2023 Deriving Dependency Graphs from Abstract Argumentation Frameworks • Strategic Argumentation: disclose those right arguments capable of persuading the other player Example 1 10
- 29. AI^3@AIxIA 2023 Deriving Dependency Graphs from Abstract Argumentation Frameworks • Strategic Argumentation: disclose those right arguments capable of persuading the other player Example 1 10
- 30. AI^3@AIxIA 2023 Deriving Dependency Graphs from Abstract Argumentation Frameworks • Strategic Argumentation: disclose those right arguments capable of persuading the other player Example 1 10
- 31. AI^3@AIxIA 2023 Deriving Dependency Graphs from Abstract Argumentation Frameworks Example 1 • Initial shared AF: G = <{Cinema, Home}, {(Cinema, Home), (Home, Cinema)}> • Initial private AF: P = <{Videogames, AngryWife, Cooking, Cinema, Home}, {(Cinema, Home), (Home, Cinema), (Videogames, Cinema), (AngryWife, Videogames), (Cooking, Cinema)}> • new {Videogames, AngryWife, Cooking} in T(Videogames)P, where T(x) = add({x}, {}) -> (c-test1(Home, in, gde) -> add({Play!}, {}) -> success) +P (true -> add({Cook!}, {}) -> success) 11
- 32. AI^3@AIxIA 2023 Deriving Dependency Graphs from Abstract Argumentation Frameworks Example 2 • Decision Making with Privacy preserved: only disclose necessary information needed to reach the desired outcome 12
- 33. AI^3@AIxIA 2023 Deriving Dependency Graphs from Abstract Argumentation Frameworks Example 2 • Decision Making with Privacy preserved: only disclose necessary information needed to reach the desired outcome 12
- 34. AI^3@AIxIA 2023 Deriving Dependency Graphs from Abstract Argumentation Frameworks Example 2 • Decision Making with Privacy preserved: only disclose necessary information needed to reach the desired outcome 12
- 35. AI^3@AIxIA 2023 Deriving Dependency Graphs from Abstract Argumentation Frameworks Example 2 • Decision Making with Privacy preserved: only disclose necessary information needed to reach the desired outcome 12
- 36. AI^3@AIxIA 2023 Deriving Dependency Graphs from Abstract Argumentation Frameworks Example 2 • Decision Making with Privacy preserved: only disclose necessary information needed to reach the desired outcome 12 new {likesOutdoors … ?
- 37. AI^3@AIxIA 2023 Deriving Dependency Graphs from Abstract Argumentation Frameworks Example 2 • Decision Making with Privacy preserved: only disclose necessary information needed to reach the desired outcome 12 new {likesOutdoors … new {likesOutdoors, likesHiking} … ?
- 38. AI^3@AIxIA 2023 Deriving Dependency Graphs from Abstract Argumentation Frameworks Example 2 • Decision Making with Privacy preserved: only disclose necessary information needed to reach the desired outcome 12 new {likesOutdoors … new {likesOutdoors, likesHiking} … ?
- 39. AI^3@AIxIA 2023 Deriving Dependency Graphs from Abstract Argumentation Frameworks Conclusion • Agents use local stores to enhance the reasoning with private information • Optional timeouts to make operators more flexible ‣ True concurrency ‣ Interleaving • In the paper: syntactic sugar that simplifies the modelling of complex situations 13
- 40. AI^3@AIxIA 2023 Deriving Dependency Graphs from Abstract Argumentation Frameworks Future Work • Integrate the Interleaving Semantics • Endow agents with a notion of ownership ‣ establish which actions can be performed on arguments • Introduce tcla constructs to identify the best sequence of actions to perform (winning strategy) • Model chatbot interactions and explanation activities 14
- 41. On the Role of Local Arguments in the (Timed) Concurrent Language for Argumentation Stefano Bistarelli, Maria Chiara Meo and Carlo Taticchi Thank you for your attention!