feedback phenomena

Terms from Artificial Intelligence: humans at the heart of algorithms

Feedback is critical in many dynamical systems with complex interactiosn between different components. If the flows of informationor effect are all in one direction as a pipeline it is possible to consider each in turn as separate inout–output systems. Hwever, if there is any feedback, where there is a loop of influnces (e.g. A affects B affects C affects C), then the types of behaviour tat can emerge are mpre complex, as effectieby the futyre behaviour of a component is affected by its own past behaviour through the loop of effects. In in cases of negative feedback, where the feedback to a component (say A) pushes it in the opposte direction to its recent changes, will typically lead to a {[stable equilibrium}} or homeostasis. However, postuve feedback where small changes lead to feedbacjk pushing A in the sam edircetion of its change, well tend to lead to runaway behaviour with large swings or exponential growth. Systems are often designed using negative feedback using closed-loop control systems

Used in Chap. 7: page 96; Chap. 23: page 368

Also known as feedback effects