Demonstrations

Should you buy the silver? Watch the brain weigh subjective value against market price, accumulating evidence toward a BUY or DON’T BUY boundary.

How does the brain decide what we perceive? See how an evidence accumulator integrates noisy sensory signals until committing to a choice.

An agent discovers a restaurant’s true quality through noisy feedback, updating its estimate one meal at a time. Watch prediction errors shrink as learning converges.

Prediction errors shift backward from reward to the earliest predictive cue over learning — the exact pattern seen in dopamine neuron recordings.

Watch values propagate backward from a goal as an agent walks Home → Park → Street → Coffee Shop. See how discount factor shapes the value gradient.

A 3-layer neural network learns to classify fruit by sweetness and redness. Watch the error ripple backward through the network via the chain rule, nudging every weight toward a better answer.

Press a lever under four different reinforcement schedules and watch the classic cumulative records emerge — from the steady slot-machine pull of variable ratio to the scalloped waiting pattern of fixed interval.

Two slot machines with hidden payoff rates. Should you exploit the best arm or explore? Temperature τ controls the balance in this Boltzmann policy demo.

The Critic learns state value V, the Actor learns action preferences m(a), both taught by a shared dopamine-like prediction error δ. Compare learning rates for each system.

Study a list of words, then complete word stems like MOT___. Does prior exposure bias what comes to mind? Experience the priming effect — a hallmark of implicit memory.

Guide a rat through a three-armed maze and watch a hippocampal place cell fire. Rotate the maze and a visual cue card to discover that place fields are anchored to landmarks, not absolute position.