UC BERKELEY
EECS technical reports
TECHNICAL REPORTS


EECS-2011-163.pdf
Conditions of Use

Archive Home Page

A Sparse Coding Method for Specification Mining and Error Localization

Authors:
Li, Wenchao
Seshia, Sanjit A.
Technical Report Identifier: EECS-2011-163
December 25, 2011
EECS-2011-163.pdf

Abstract: Formal specifications play a central role in the design, verification, and debugging of systems. We consider the problem of mining specifications from simulation or execution traces of reactive systems with a special focus on digital circuits. We propose a novel sparse coding method that can extract specifications in the form of a set of basis subtraces. For a set of finite subtraces each of length p, we introduce the sparse Boolean basis problem as the decomposition of each subtrace into a Boolean combination of only a small number of basis subtraces of the same dimension. The contributions of this paper are (1) we formally define the sparse Boolean basis problem and propose a graph-based algorithm to solve it; (2) we demonstrate that we can mine useful specifications using our sparse coding method; (3) we show that the computed bases can be used to do simultaneous error localization and error explanation in a setting that is especially applicable to post-silicon debugging.