Details
Poster
Presenter(s)
![Muhammad Gad Headshot](https://confcats-catavault.s3.amazonaws.com/CATAVault/ieeecass/master/files/styles/cc_user_photo/s3/user-pictures/20511.jpg?h=ff07c561&itok=tjaofQaR)
Display Name
Muhammad Gad
- Affiliation
-
AffiliationGerman University in Cairo
- Country
-
CountryEgypt
Abstract
A ML-based approach for automating and improving the process of sequential circuits\' verification is proposed. The method employs adaptive neural network with shortest path algorithm to generate a directed sequence of tests (instructions) to improve overall coverage. The proposed method achieve coverage closure when used to verify a quad-core cache design implementing MESI protocol. The employed adaptive neural network also overcomes the problem of sensitivity of learning algorithms to the size and quality of the initial training set.