Intention recognition using conceptual fuzzy sets

In the man-machine interface, the user´s goal is transmitted to the system, the system understands the goal, and the processing is carried out. The status of the system and the processing result are transmitted to the user. Therefore, to realize this communication, it is necessary to recognize a user´s intention from his or her words. This intention recognition raises the computer system to the information processing level of a human. A conceptual fuzzy set (CFS) represents the context-dependent meaning of a vague concept. Multilayered reasoning (MLR) based on CFS can provide abstract and concrete representation simultaneously, and can handle top-down and bottom-up processing simultaneously. The authors apply MLR based on CFS to intention recognition. Goal recognition for a drawing tool is used as a concrete example. The multistructured knowledge for goal recognition is represented using CFS, and is inductively constructed using the Hebbian learning law