The Validation of Graph Model-Based, Gate Level Low-Dimensional Feature Data for Machine Learning Applications

04/05/2021
by   Aneesh Balakrishnan, et al.
0

As an alternative to traditional fault injection-based methodologies and to explore the applicability of modern machine learning algorithms in the field of reliability engineering, this paper proposes a systemic framework that explores gate-level netlist circuit abstractions to extract and exploit relevant feature representations in a low-dimensional vector space. A scalable feature learning method on a graphical domain called node2vec algorithm had been utilized for efficiently extracting structural features of the netlist, providing a valuable database to exercise a selection of machine learning (ML) or deep learning (DL) algorithms aiming at predicting fault propagation metrics. The current work proposes to model the gate-level netlist as a Probabilistic Bayesian Graph (PGB) in the form of a Graph Modeling Language (GML) format. To accomplish this goal, a Verilog Procedural Interface (VPI) library linked to standard simulation tools has been built to map gate-level netlist into the graph model. The extracted features have been used for predicting the Functional Derating (FDR) factors of individual flip-flops of a given circuit through Support Vector Machine (SVM) and Deep Neural Network (DNN) algorithms. The results of the approach have been compared against data obtained through first-principles approaches. The whole experiment was implemented on the features extracted from the 10-Gigabit Ethernet MAC IEEE 802.3 standard circuit.

READ FULL TEXT

page 2

page 6

research
04/05/2021

Composing Graph Theory and Deep Neural Networks to Evaluate SEU Type Soft Error Effects

Rapidly shrinking technology node and voltage scaling increase the susce...
research
04/05/2021

Modeling Gate-Level Abstraction Hierarchy Using Graph Convolutional Neural Networks to Predict Functional De-Rating Factors

The paper is proposing a methodology for modeling a gate-level netlist u...
research
02/24/2018

SAT-based Reverse Engineering of Gate-Level Schematics using Fault Injection and Probing

Gate camouflaging is a known security enhancement technique that tries t...
research
02/03/2021

Machine Learning-assisted Fault Injection

Fault Injection (FI) is a method for system validation and verification ...
research
08/05/2021

Extending the Graph Formalism to Higher-Order Gates

We present an algorithm for efficiently simulating a quantum circuit in ...
research
01/06/2022

Skip Vectors for RDF Data: Extraction Based on the Complexity of Feature Patterns

The Resource Description Framework (RDF) is a framework for describing m...
research
05/30/2022

End-to-End Topology-Aware Machine Learning for Power System Reliability Assessment

Conventional power system reliability suffers from the long run time of ...

Please sign up or login with your details

Forgot password? Click here to reset