1 / 18

Learning Description From Examples

Learning Description From Examples. Learning by Analyzing Differences: the Winston Algorithm. Artificial Intelligence Learning by Analyzing Differences. L. Manevitz. Procedure W Let description  First Sample. For all samples: If near miss use SPECIALIZE. If example use GENERALIZE.

bcaroline
Download Presentation

Learning Description From Examples

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. Learning Description From Examples Learning by Analyzing Differences: the Winston Algorithm

  2. Artificial IntelligenceLearning by Analyzing Differences L. Manevitz

  3. Procedure W • Let description  First Sample. • For all samples: • If near miss use SPECIALIZE. • If example use GENERALIZE.

  4. “Near-Miss” • Need to find single important difference from example to description • If cant identify the difference, then skip example

  5. Example Arch Near miss Near miss Arch

  6. A B C Wedge Brick a Structural Descriptions has-part has-part supported-by Is-a Is-a

  7. C A B D Brick Brick Brick b Structural Descriptions cont. has-part has-part has-part supported-by supported-by Is-a left-of Is-a Is-a right-of Does-not-marry

  8. C A B D Brick Brick Wedge c Structural Descriptions cont. has-part has-part has-part supported-by supported-by Is-a left-of Is-a Is-a right-of Does-not-marry

  9. must-support must-support left-of Arch a b c left-of Near miss Figure 2 support support left-of Arch

  10. must-support must-support Arch left-of must-not-touch must-not-touch c b a Figure 3 must-support must-support must-support must-support Arch Near miss left-of left-of touch touch

  11. Is-a Brick Block must-support must-support Arch left-of must-not-touch must-not-touch a Figure 4

  12. Is-a Wedge Block must-support must-support Arch left-of must-not-touch b must-not-touch Figure 4 cont.

  13. must-be-a Block must-support must-support Arch left-of must-not-touch c must-not-touch Figure 4 cont.

  14. Specialize • Match the evolving model to the sample to establish correspondences among parts.

  15. Specialize cont. • Determine whether there is a single, most important difference between the evolving model and the near miss: • If there is a single, most important difference , determine whether the evolving model or the near miss has a link that is not in the other: • If the evolving model has a link that is not in the near miss, use the require-link heuristic. • If the near miss has a link that is not in the model, use the forbid-link heuristic. • Otherwise ignore the sample.

  16. Generalize • Match the evolving model to the sample to establish correspondences among parts. • For each difference, determine the difference type: • If the difference is that the link points to a different class in the evolving model from the class the link points to in the sample, determine if the classes are part of a classification tree:

  17. Generalize cont. • If the classes are part of a classification tree, use the climb-tree heuristic. • If the classes form an exhaustive set, use the drop-link heuristic. • Otherwise, use the enlarge-set heuristic. • If the difference is that a link is missing in either the evolving model or the example, use the drop-link heuristic. • If the difference is that different numbers, or an interval and a number outside the interval, are involved, use the close-interval heuristic. • Otherwise ignore the difference.

  18. The Heuristics • The require-link heuristic. • The forbid-link heuristic. • The climb-tree heuristic. • The enlarge-set heuristic. • The drop-link heuristic. • The close-interval heuristic.

More Related