Do you know how training algorithms work on a simple basis? Like the ones your car uses to adapt fuel and air flow with a feedback sensor? They are super simple but it's helpful for understanding how your brain trains to make probability estimates.

So in engineering we call these control systems. Your body is full of them, probably how we figured out we could build them.

They involve inputs, outputs, signals, and feedback

Your thermostat is a very simple example. You have an electrical circuit that turns your heater on or off based on the feedback of a temperature senser and a set ideal temperature.

You set the temperature at 72. If the temperature is below that, the circuit is closed and the furnace inputs heat into the environment. The thermometer measures the temperature and when temperature reaches 72 it opens the circuit and the furnace switches off.

Then heat leaves the environment through transfer to exterior or ground or whatever. When the temp again drops below 72 the circuit closes.

For this system we don't have to measure the heat entering or leaving because temperature is easy to measure.

Now your car engine has some more sophisticated models meant to adapt to different environmental conditions, fuel quality, wear and tear on the engine etc.

So simplified it inputs air and fuel into the combustion chamber...

And measures the result with a sensor that measures residual oxygen and recalculates the inputs based on that. But...

It learns. If the air fuel input is missing the mark over time, it adjusts them permanently to a different level. This require measurement of inputs.

Your brain has cells like this! They measure inputs and results and adjust over time to accommodate environmental factors!

And some of these cells influence our behavior and decisionmaking!

You are a car, your perceptions are feedback, your behavior is an input, your brain is building in adjustments.

Wait did I lose you when I started talking about cars? Sorry I do that.