Need is All You Need: Homeostatic Neural Networks Adapt to Concept Shift

In living organisms, homeostasis is the natural regulation of internal states aimed at maintaining conditions compatible with life. Here, we introduce an artificial neural network that incorporates some homeostatic features. Its own computing substrate is placed in a needful and vulnerable relation to the very objects over which it computes. For example, MNIST digits may cause excitatory or inhibitory effects upon the homeostatic network that classifies them, by altering the network’s learning rate. Accurate recognition is desirable to the agent itself because it guides decisions to up- or down-regulate its internal states and functionality. Counter-intuitively, the addition of vulnerability to a learner can confer some benefits. Homeostatic learners are more adaptive under conditions of concept shift, in which the relationships between labels and data change over time. The greatest advantages are obtained under the highest rates of shift. Homeostatic learners are also resilient to second-order shift, or environments with changing rates of concept shift.