Home

About

Advanced Search

Browse by Discipline

Scientific Societies

E-print Alerts

Add E-prints

E-print Network
FAQHELPSITE MAPCONTACT US


  Advanced Search  

 
Adaptive Memoization Umut A. Acar Guy E. Blelloch Robert Harper
 

Summary: Adaptive Memoization
Umut A. Acar Guy E. Blelloch Robert Harper
Computer Science Department
Carnegie Mellon University
Pittsburgh, PA 15213
{umut,blelloch,rwh}@cs.cmu.edu
Abstract
Memoization may be viewed as a mechanism for re-using
a computation--if a function is re-applied to the same ar-
gument we may re-use the previous computation to deter-
mine the result, rather than perform it again. Conventional
memoization is accurate in the sense that it only permits
re-use when a computation will be precisely the same as
one that has already been performed. But in many cases a
computation may be largely, though not entirely, the same
as one that has been previously carried out for a slightly
different input. The previous computation may be re-used
by permitting inaccurate memoization, and restoring accu-
rancy by adapting the result to the variant input. This tech-
nique, which we call adaptive memoization, greatly increases

  

Source: Acar, Umut - Programming Languages and Systems Group, Max-Planck Institute for Software Systems

 

Collections: Computer Technologies and Information Sciences