LEARNING (Competitive)
Charles François (2004). LEARNING (Competitive), International Encyclopedia of Systems and Cybernetics, 2(1): 1881.
| Collection | International Encyclopedia of Systems and Cybernetics |
|---|---|
| Year | 2004 |
| Vol. (num.) | 2(1) |
| ID | ◀ 1881 ▶ |
| Object type | Discipline oriented, Methodology or model |
In parallel processing networks , “an unsupervised learning procedure in which the network is presented with a series of input patterns and must discover regularities in those patterns that can be used to divide them into clusters of similar patterns ” (Adapted from W. BECHTEL & A. ABRAHAMSEN, 1991, p.304)
The authors add that input patterns are selected through “inhibitory connections among thedetector units to assure that just one unit will ”win“ the competition for a particular input pattern ” (p.305).
In this way, inputs become classified in categories .
- “In more complex systems there may be multiple nets of detector units, or intermediate layers of units” (p. 305)
Interesting heuristic analogies can be drawn with the ways human brains are learning.