The First Stage of Prediction – Learning
When a child learns something new and newish, they look around and ask questions. The answers and the words used are simple answers to what may not always be simple.
For example, if they look at a dog, what do they see? Why is it a dog? Why is it called a dog?
A few of those answers might be:
- A dog barks (so does a seal, but would a young child see a seal?)
- A dog growls (so does a bear, but would they see one?)
- A dog eats meat (a dog will eat a lot of what you put in front of it. It’s omnivorous – within reason, it is a dog
Anything else is pretty much breed defined:
- Is it big or small?
- Furry or bald?
- What colour is a dog’s fur?
Learning is basic and experiential, and the person who answers the questions is limiting answers based on what they think the child needs to know and what will quickly explain it. And what will protect the child.
Will a dog hurt me? That depends:
- How is the dog controlled?
- How is the child treating the dog?
And from this the child develops their schema for what a dog is. And adds new info as they run into dogs. That look like big rats, hairy donkeys and bears. That are timid, frightened creatures or really aggressive, terrifying creatures.
And they learn how to predict the dog’s behaviour. How to trust the dog.
And most of these rules and methods also are applicable to teaching a computer what something new is to them. You supply info they need to understand what the new thing is in terms they can understand.
First comes understanding then comes prediction. To what degree though? Do they want to record, define, understand, forecast it’s behaviour, or control the new thing?
And was the info accurate. Garbage in, garbage out.
But for either child or computer, the info is basic and repetitive till they build enough schema to start creating for themselves. A child will, but will a computer?