The redecentralization of the web introduces new challenges on trusting data from other sources due to many unknown or even hidden parties. An application working trustworthy in a decentralized web must evaluate trust and take trustaware decisions autonomously without relying on a centralized infrastructure. This autonomy and the huge amount of available applications necessitates the web to be modelled as an open dynamic Multi-Agent System (MAS). To evaluate the trust of web agents, the most suitable trust models need to be identified and used. Despite the various trust models proposed in the literature for evaluating a web agent’s trust, the examination of them with different scenarios and configurations is not trivial. To address these challenges, we initiated aTLAS, a Trust Laboratory of Multi-Agent Systems which is a web-based wizard testbed for researchers and web engineers to evaluate trust models systematically. aTLAS will enable future research regarding trust evaluations in a decentralized web.
The aTLAS project intends to examine trust for a redecentralization of the web. It enables a broad comparison of trust mechanics, scales and models from the literature within the current state of the art. Therefore, it runs and evaluates multi-agent system scenarios, which are defined beforehand. As the redencentralization of the web necessitates it to be modeled as a open dynamic multi-agent system, such a laboratory can support the current situation where a comparision of trust approaches for a decentralized web has to be done manually with a high effort.
ATAC 2021 @ ISWC 2021 Paper (ATAC'21 Challenge Winner)
For the prototype check out the aTLAS Online Prototype and aTLAS Code Repository