Pitfalls in Practical Open Multi Agent Argumentation Systems: Malicious Argumentation

摘要: When an interaction mechanism such as argumentation is considered for use in open multi agent domains, E-Commerce or other business applications, it necessary to consider the possibility of agents performing malicious actions. Common themes when studying actions communication protocols are that withholding information and misrepresenting information. In argumentation, however, a complex underlying formal logic allows another type action: introduction superfluous complexity into information, designed overwhelm reasoning capacity agent. We examine strategy systems based on exploiting order manipulate outcome argument acceptability evaluation. Further, we briefly discuss general problem defensive strategies against this inherent difficulty detecting occurrences it.