Alberto Pettorossi

Title: Alberto Pettorossi

Main Research Question: How can we improve the efficiency of logic programs through specialization and reduction of nondeterminism?

Methodology: The authors propose a set of transformation rules that extend those used for partial evaluation, and a strategy for guiding their application to derive specialized, highly efficient programs. They use mode information to guide the unfolding process and reduce nondeterminism.

Results: The authors demonstrate that their technique can automatically derive very efficient matching programs and parsers for regular languages, which could not have been done by previously known partial evaluation techniques. They show that their approach can lead to efficiency improvements that are sometimes exponential, due to the reduction of nondeterminism and the fact that computations performed in different branches of the computation trees are performed within single branches.

Implications: The technique proposed by Pettorossi et al. can lead to significant improvements in the efficiency of logic programs, particularly in reducing nondeterminism and improving program specialization. This can have practical implications in various fields, such as natural language processing, pattern recognition, and automated reasoning.

Link to Article: https://arxiv.org/abs/0311044v1 Authors: arXiv ID: 0311044v1