| 
|
QuoteRef: lenaDB8_1984
topics >
all references >
ThesaHelp:
references i-l
Topic:
heuristic-based systems
Topic:
attribute-value pairs as information
Topic:
knowledge representation by frames
Topic:
self-regulating systems
Reference
Lenat, D.B., Brown, J.S., "Why AM and EURISKO appear to work", Artificial Intelligence, 23, 3, August 1984, pp. 269-294. Google
273 ;;Quote: AM and EURISKO do not mutate the meanings of concepts, they mutate the structural form of a concept in some representation
275 ;;Quote: an EURISKO heuristic is 20-30 slots with a line or two of code, an atom, or a short list
276 ;;Quote: in non-mathematical domains, EURISKO needs a rich set of slots for the domain and a frame of information about that type of slot
289 ;;Quote: AM found many interesting results (not all known to Lenat) with multiple discoveries per heuristic, and many heuristics per discovery
290 ;;Quote: for AM, started with a large number of concepts, slot definitions, an some initial values; only 1% of final knowledge added later
291 ;;Quote: AM exploited the natural tie between LISP and math by syntactically mutating small LISP programs
292 ;;Quote: AM and EURISKO are like perceptrons, i.e., societies of evolving, self-organizing, symbolic knowledge structures
ThesaHelp: references i-l (342 items)
Topic: heuristic-based systems (35 items)
Topic: attribute-value pairs as information (53 items)
Topic: knowledge representation by frames (18 items)
Topic: self-regulating systems (23 items)