Publikation

Contract-based Data-sharing for AI-based Decision Making on the Web

Wolfgang Maaß

In: 55th Hawaii International Conference on System Sciences. Hawaii International Conference on System Sciences (HICSS-2022) January 4-7 Hawaii Hawaii United States Seiten 1997-2006 Springer 2022.

Zusammenfassung

The data-driven economy on the World Wide Web are based on coordination mechanisms for the exchange of and AI-based processing of data sets created by independent actors. Architectures for data-driven economic systems currently focus on the exchange of datasets and leave the processing of datasets to background mechanisms. Without binding commitments, agents in data sharing situations favor the ”no data sharing” strategy. We present a broker framework that facilitates contract-based exchange, trading, and processing of data sets between agents under conditions of ”lack of trust”. Electronic contracts guarantee ownership and control of data datasets, the execution of defined data analysis tasks based on AI models, and the sharing of results according to contractual commitments. A technical architecture (TUCANA) is presented that provides a federated data economic ecosystem including a broker framework. We present an application based on an implementation of TUCANA.

Projekte

HICSS_Maas_2022.pdf (pdf, 941 KB )

Deutsches Forschungszentrum für Künstliche Intelligenz
German Research Center for Artificial Intelligence