NISPS Playground

NISPS Playground

Train a neural network to map joystick positions to visual parameters using interactive machine learning.

Examples Mode

1. Move the joystick to a position
2. Drag the parameter bars to set desired visual output
3. Press Add Example to save this mapping
4. Repeat for different joystick positions
5. Press Train — the network learns to interpolate between your examples
6. Move the joystick — visuals respond through the learned mapping

RL Feedback Mode

1. Move the joystick around — the network produces different visual outputs
2. See something you like? Press + (thumbs up)
3. Don't like it? Press − (thumbs down) to explore more
4. Double-click the joystick to toggle Follow mode (no hold needed)
5. In Follow mode: keyboard 2 = thumbs up, 1 = thumbs down
6. With a gamepad: RB = thumbs up, LB = thumbs down
7. The network learns from your preferences over time

Tips

Try the preset buttons above the visuals for instant demos. Your training data auto-saves to your browser.

Tap anywhere outside to close