Think Together · Start with a Short Story
The girl says, “Keep the method you learned.” What does the robot save? Not a tiny teacher, but a computing structure and its parameters.
Lay Out the Process
- StructureDefines how calculations work
- ParametersValues adjusted during training
- InputNew data
- OutputA prediction or generated result
A model is a computational representation that maps inputs to outputs.
Training changes the parameters to make the model more suitable for a task. You can imagine parameters as many small dials that control calculations, while the structure determines how the dials connect.
When using a model, new input goes in, and the model calculates a result using its existing structure and parameters.
Now You Try
If the same structure gets a different set of parameters after training, could its output change?
Ready? Open this to see the explanation
Yes. The same structure does not mean it learned the same relationships. Parameters affect the calculation results.
One Idea to Take Away
A model is a computational representation that maps inputs to outputs.
Try covering the diagram above and explaining each step to someone in your family. If one step is hard to explain, return to that card and look again.