GL 3 Learning rules
The GL 3 learns new rules while the sleeping, because that activity takes a considerable time. It can learn new rules in several ways:
- One way is, that it observes elementary rules and when it finds a rule where the future situation (FS) is equal to the present situation (sit1) of another rule, it puts them together, making a new rule that goes from the sit1 of the first rule to the FS of the second rule. It stores the intermediate common situation in the IntSits part of the new rule. In that way more and more rules are connected.
- Another way is learning from errors. When the FS reached was not the one expected, due to an incomplete sit or action side effects, then the GL 3 creates a new rule with the FS really reached.
- Still another is from the experiences stored in memory. This is identical to the activity that the old General Learner did, only that this time not only external rules are involved, but also internal rules.
- Learning by copying: The activity of the person is stored as rules and so new rules are learned.
- Learning by curiosity: When there are parts of a sit that have few applicable rules, make rules with the abstract of the concept and with each concrete example of the concept.
- In future: Make also new rules with mirror objects. change all +x to -x and all +y to -y.
- Use the General Pattern Finder to create new rules from those in memory by finding new relationships (new patterns) within any of the situation parts.
See General Pattern Finder
(Enter for continuos reading, like a book)..
Teaching
Besides learning by itself, there is the lerarning from experiences created by the teacher. The teacher creates exterior and interior situations & change of situations (by actions) for learning from experience.
The teacher can write on the blackboard. The GL 3 senses this and learns the corresponding rules.
When teaching, start with the easy and only later do the difficult. When learning mathematics, we start with adding and subtracting and only later do differential equations. The reason is that for doing differential equations you need to know adding and subtracting, but for adding you don't need to know differential equations.
Plan the learning top down, but do the learning, step by step, bottom up. For each test case, the learning of external capabilities should be done at the same time as the learning of those internal capabilities needed for the external capabilities.
For continuos reading, like a book - continue
here.
Jump to the top of this document /
Present Investigation /
E-Book Contents.