Interdependence: A mathematical approach to the autonomy of human-machine systems

Open Access
Conference Proceedings
Authors: William LawlessDonald Sofge

Abstract: We update our theory of interdependence for autonomous human-machine teams operating in open systems (A-HMT-S). In closed systems, desired outcomes can be easily obtained with rational models (e.g., game theory); there, uncertainty can only be studied as part of a system’s internal com-plexity. In hindsight, the problems with closed system models are obvious: they are fragile, hard to replicate, and not generalizable, the latter being the fatal flaw for autonomous human-machine teams and systems. Surprisingly, no amount or aggregation of data from individuals can be recombined to replicate social data. In contrast, with open systems, interdependence theo-ry is state dependent, reactive to every situation and change, especially the environmental and social uncertainty caused by competition or conflict. More important, in contrast to social science’s reliance on the independent and identically distributed (i.i.d.) data derived from individuals, interde-pendence theory is generalizable. But before we start, we acknowledge that machine learning is a closed system model, context dependent, and that ex-isting artificial intelligence (AI) models are insufficient to produce auton-omy today. Thus, we built a mathematical model based on first principles around interdependence and applicable to intelligent autonomous teams of any sort. With our model of interdependence, among the results we have found: reactiveness to bistable information requires intelligence, and boundaries as a barrier to impede its unwanted flow; independent infor-mation cannot replicate teammate dependence, effects, nor performance; in-terdependence creates tradeoffs between the structure and performance of autonomous systems, that, as byproducts, affords metrics, deception, sup-pression, and vulnerability, the latter being a new field of research that we have discovered and that is the motivation for innovation, mergers and ac-quisitions. We close with a brief review of future research opportunities.

Keywords: autonomy, human-machine teams and systems, interdependence

DOI: 10.54941/ahfe1002311

Cite this paper: