Adaptive AI Model Achieves Synaptic Stability and Plasticity

Researchers developed an adaptive AI model grounded in the TSKI-4.1 theoretical framework. This model used original formulas, principles, and algorithms. A technical specification was prepared, describing the strict order of tick stages, a mirror synapse model, and rules of plasticity and stabilization according to TSKI-4.1. The key goals of this simulation were to demonstrate the model's adaptive properties and its capacity to stabilize synaptic plasticity within a group of neurons.

During the simulation, the model's core capabilities were confirmed. As lead author Atorin notes in the paper, "During the simulation, the following were confirmed: the model’s adaptivity; its ability to stabilize the number of synapses in a group under synaptic plasticity; and its capacity to maintain the final number of synapses after active restructuring of connections." These confirmed capabilities demonstrate the model’s adaptivity, its ability to stabilize the number of synapses in a group under synaptic plasticity, and its capacity to maintain the final number of synapses after active restructuring of connections.

The study also analyzed the roles of feedback loops and the inhibitory neuron.