=========================================================
P L A N N E D   L E A R N I N G   C A P A B I L I T I E S
=========================================================

This file is a description of future learning mechanisms.


-= Learning of default values =-

If the user does not provide full information about room or event
properties, the system uses defaults. For example, if the user asks to
allocate a room for an invited talk, the system may need to make
assumptions about the room size, need for a projector, and so on. The
purpose of the "default learning" is to create appropriate defaults by
analyzing known values in the current world model and other available
world models. The system should update the learned defaults as it
receives more information during the "war games."


-= Learning of default rules =-

The system should learn conditions of default rules, which represent
appropriate contexts for default assumptions.


-= Learning of relevant questions =-

This learning mechanism identifies the patterns that distinguish
useful questions from less useful ones. It analyzes the questions
generated during "war games" and learns which of them turn out useful.
For example, the system may learn that it should always find the exact
size of large rooms and check the availability of projectors. After
the system learns these patterns, it can generate questions by a
"static analysis" of the available resources.