We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.AI

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Artificial Intelligence

Title: Corruption and Audit in Strategic Argumentation

Abstract: Strategic argumentation provides a simple model of disputation and negotiation among agents. Although agents might be expected to act in our best interests, there is little that enforces such behaviour. (Maher, 2016) introduced a model of corruption and resistance to corruption within strategic argumentation. In this paper we identify corrupt behaviours that are not detected in that formulation. We strengthen the model to detect such behaviours, and show that, under the strengthened model, all the strategic aims in (Maher, 2016) are resistant to corruption.
Comments: Reasoning Research Institute technical report, 2017
Subjects: Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR); Multiagent Systems (cs.MA)
MSC classes: 68T30, 68T35, 68T42
ACM classes: I.2.4; F.2.2
Cite as: arXiv:2008.13115 [cs.AI]
  (or arXiv:2008.13115v1 [cs.AI] for this version)

Submission history

From: Michael Maher [view email]
[v1] Sun, 30 Aug 2020 08:08:26 GMT (207kb,D)

Link back to: arXiv, form interface, contact.