[ CogSci Summaries home | UP | email ]

Narayanan, N. H., Suwa, M. & Motoda, H. (1994). How things appear to work: Predicting behaviors from device diagrams. Proceedings of the 12th National Conference on Artificial Intelligence, AAAI Press, pp. 1161-1167.

  author = 	 {Narayanan, Hari N. and 
                  Suwa, M. and 
		  Motoda, H.},
  title = 	 {How things appear to work: Predicting behaviors from device diagrams},
  booktitle = 	 {Proceedings of the
12th National Conference on Artificial Intelligence},
  OPTcrossref =  {},
  OPTkey = 	 {},
  OPTpages = 	 {1161--1167},
  OPTyear = 	 {1994},
  OPTeditor = 	 {},
  OPTvolume = 	 {2},
  OPTnumber = 	 {},
  OPTseries = 	 {},
  OPTaddress = 	 {},
  OPTmonth = 	 {July},
  OPTorganization = {AAAI},
  OPTpublisher = {AAAI Press},
  OPTnote = 	 {},
  OPTannote = 	 {}

Author of the summary: Jim Davies, 2004, jim@jimdavies.org

Cite this paper for:

The agent's task is to "predict the operation of the device by hypothesizing behaviors of its compoents, given a labeled schematic diagram of the device showing the spatial configuration of its components and an initial condition or behavior."[1161]

Uses visual and conceptual reasoning. Experiment: Ran 5 subjects in a protocol experiment. Findings are
1. the diagram facilitated memory cueing.
2. the diagram supported mental simulation which allowed the reasoner to predict effects of simulated behaviors. [1162]

The experiment inspired the cognitive model, which uses a diagram and domain-knowledge to predict the behavior of the diagrammed system. The visual representation contains "diagram frames" and array representations (like Glasow 1992). The array representation contains knowledge of what is contained in each pixel. The user provides higher-level descriptions of what these pixels represent. At present they are line segments and locations. [1164]

Visual reasoning:

Reasoning switches to diagrammatic reasoning when either 1. the inference is a hypothesis about spatial behavior and the consequences with respect to other components need to be determined, or 2. it is a hypothesis about a spatial behavior whose immediate consequences have already been determined. [1166]

This paper uses a combination of visual and non-visual knowledge to reason about a diagrammed situation. It is valuable because it shows when visual knowledge and reasoning are useful.

Summary author's notes:

Back to the Cognitive Science Summaries homepage
Cognitive Science Summaries Webmaster:
JimDavies (jim@jimdavies.org)
Last modified: Tue May 13 10:28:57 EDT 2003