Computational Emergence and Emergent Computation: A Duality in Research on Artificial Collective Behaviors
We elaborate on computational emergence (CE), understood as the emergent acquisition of specific abilities from specific forms of computation, such as artificial neural networks and cascades of rule iterations found in cellular automata. CE leads to the acquisition of properties such as learning abilities, morphological pattern formation, and coherence, and arises from computational mechanisms. We also elaborate on emergent computation (EC), understood as the emergent acquisition of computational abilities by communities of phenomenologically interacting agents, potentially through appropriate interlinkages among them, as in emerging networks. Processes of interaction are understood generically as forms of mutually active interdependence, which can be modeled as self-generated networks. EC arises from phenomenological mechanisms of interaction among agents and leads to the acquisition of properties such as coherent behaviors, resilience, robustness, and collective intelligence. The reason for distinguishing between these two types of emergence is that doing so may open new approaches to modeling collective behavior, especially in artificial ones, such as swarms of unmanned aerial vehicles (UAVs), where introducing parametric and structural changes is more feasible. Combining the two approaches—(a) phenomenological, networked EC arising from populations of interacting (b) in turn computationally emergent agents—allows the consideration of research directions such as identifying relationships between combinations of CE and emergently acquired computational properties within the conceptual frameworks of networked neural networks and intersected neural networks, i.e., networks that share neurons. Such research directions are expected to enable approaches for influencing collective behaviors and complex systems in a non-invasive way, including swarms of UAVs (or drones), autonomous cyborg swarms, and coherent communities of artificial devices equipped with sensors, edge artificial intelligence, and secure communications. We consider the mesoscopic nature of complexity in collective behaviors as a continuous negotiation between these two forms of emergence, with EC playing a macroscopic role and CE a microscopic role. We conclude that this general framework relates to the concept of “The Middle Way” in physics by focusing on what occurs “in between” systems (such as between intersecting neural networks and their dynamic networking) and within transient spaces where non-invasive intervention may be possible and appropriate for guiding, modifying, and inducing changes in complex emergent systems.