In the Artificial Intelligence community, abstraction is typically viewed as a mechanism for attacking intractable search problems. For example, the performance of a planning algorithm may be improved by using a hierarchy of abstraction spaces. Search algorithms use such abstraction techniques in order to make search control decisions.
Have the two communities missed an opportunity for cross fertilization? For example, do abstraction techniques developed in SE have any application to AI? Do AI techniques for automatically generating abstractions have any application to SE?
\bibitem{Cousot00-SARA-Panel}
P{.} Cousot.
\newblock Contribution to the panel on ``Abstractions in AI and software
engineering''.
\newblock Fourth International Symposium on Abstraction, Reformulations and
Approximation, SARA'2000, Horseshoe Bay, Texas, United States, 26--29
July 2000.
@unpublished{Cousot00-SARA-Panel,
author = {Cousot, P{.}},
title = {Contribution to the Panel on ``{A}bstractions in {AI} and Software
Engineering''},
note = {Fourth International Symposium on Abstraction, Reformulations and
Approximation, SARA'2000, Horseshoe Bay, Texas, United States,
26--29 July 2000},
}