Self-Modification

Self-modification can be seen as a form of learning. However, self-modification is distinct from learning for two specific reasons. First, learning refers to the addition of new knowledge. Self-modification may be extended to include changes to the basic architectural properties of the agent (e.g., reasoning mechanisms, knowledge structures, etc.). Second, learning is usually purposeful. Self-modification, on the other hand, may involve random or probabilistic changes that are not goal-motivated. Thus, self-modification can be viewed more as an evolutionary architectural mechanism involving mutations rather than simply a type of learning.


Press UP to go the property list.

Press HOME to return to the Table of Contents.