When solving a problem, humans often draw diagrams in order to acquire useful information such as the conditions that are not clearly specified. The importance of using diagrams for reasoning and problem solving lies not only in clarifying humans' intellectual activities, but also in developing efficient user interfaces. There are several studies in the field of problem solving systems that use diagrammatic information such as those of determining object positions in physics problems, and those of corresponding between diagrams and given problem descriptions. However, there are few studies for the processes of drawing diagrams derived from the textual description of the problem and for enumerating quantities or finding patterns. This paper proposes a method for solving arithmetic problems using diagrams, and describes our problem solver called Diagrammatic Arithmetic Problem Solver (DIPS). DIPS employs an extended semantic network as its representation of diagrams, and it successfully solves arithmetic problems by drawing diagrams from the descriptions of the problems. It also has the abilities of discovering formulas from the regularities of drawn diagrams. © 2002 Wiley Periodicals, Inc. Syst Comp Jpn, 33(6): 112–120, 2002; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/scj.1138
[1]
Gordon S. Novak.
Representations of Knowledge in a Program for Solving Physics Problems
,
1977,
IJCAI.
[2]
Gordon S. Novak,et al.
Understanding Natural Language with Diagrams
,
1990,
AAAI.
[3]
Herbert A. Simon,et al.
Why a Diagram is (Sometimes) Worth Ten Thousand Words
,
1987
.
[4]
Masaki Suwa,et al.
How Things Appear to Work: Predicting Behaviors from Device Diagrams
,
1994,
AAAI.
[5]
John Domingue,et al.
Software visualization : programming as a multimedia experience
,
1998
.
[6]
Brian V. Funt,et al.
Problem-Solving with Diagrammatic Representations
,
1980,
Artif. Intell..