Human-Machine Collective Intelligence Environment for Adaptive Decision Support
Alexander V. Smirnov, Andrew Ponomarev · 2019
Crowd computing has become an important and widely used way to solve various problems by joint effort of humans and machines. In most of the systems, leveraging elements of crowd computing, workflow or algorithm (how to split the original problem into parts, how to distribute them between human participants and how to merge the results received from different human participants) is defined as a part of system design. While it pays back in a wide range of tasks (usually simple ones), it is widely recognized that rigid workflows turn out to be very limiting when applied to complex problems (e.g., decision support). The paper proposes an approach to create flexible and adaptive decision support systems on the basis of an environment, supporting human-machine collective intelligence. The distinctive features of the proposed environment are: a) fusion of elements of collective intelligence and artificial intelligence, b) support for natural self-organization processes in the community of participants (supported by self-organization protocols, convenient for different kinds of participants of a heterogeneous human-machine system), c) interoperability of participants (both human and machine), achieved via multi-aspect ontologies, d) soft guidance in the process of self-organization.