Abstract:The globalization of the Integrated Circuit (IC) market is attracting an ever-growing number of partners, while remarkably lengthening the supply chain. Thereby, security concerns, such as those imposed by functional Reverse Engineering (RE), have become quintessential. RE leads to disclosure of confidential information to competitors, potentially enabling the theft of intellectual property. Traditional functional RE methods analyze a given gate-level netlist through employing pattern matching towards reconstructing the underlying basic blocks, and hence, reverse engineer the circuit's function. In this work, we are the first to demonstrate that applying Approximate Computing (AxC) principles to circuits significantly improves the resiliency against RE. This is attributed to the increased complexity in the underlying pattern-matching process. The resiliency remains effective even for Graph Neural Networks (GNNs) that are presently one of the most powerful state-of-the-art techniques in functional RE. Using AxC, we demonstrate a substantial reduction in GNN average classification accuracy-- from 98% to a mere 53%. To surmount the challenges introduced by AxC in RE, we propose the highly promising AppGNN platform, which enables GNNs (still being trained on exact circuits) to: (i) perform accurate classifications, and (ii) reverse engineer the circuit functionality, notwithstanding the applied approximation technique. AppGNN accomplishes this by implementing a novel graph-based node sampling approach that mimics generic approximation methodologies, requiring zero knowledge of the targeted approximation type. We perform an extensive evaluation and show that, using our method, we can improve the classification accuracy from 53% to 81% when classifying approximate adder circuits that have been generated using evolutionary algorithms, which our method is oblivious of.

Prior Knowledge Input Neural Network Method for GFET Description

Compact Modeling of Advanced Gate-All-Around Nanosheet FETs Using Artificial Neural Network

ASIC-based Gauss Error Function Module Design for Neural Network Application

Modeling Nanoscale FinFET Performance by a Neural Network Method

Compact modeling technology for the simulation of integrated circuits based on graphene field-effect transistors

GFET Lab: A Graphene Field-Effect Transistor TCAD Tool

A neural compact model based on transfer learning for organic FETs with Gaussian disorder

A New Compact Z$^{\text{2}}$-FET Model Based on Artificial Neural Network and Its Applications

A Comprehensive Graphene FET Model for Circuit Design

Artificial Neural Network Models for Metal-Ferroelectric-insulator-semiconductor Ferroelectric Tunnel Junction Memristor

Knowledge-based neural network SPICE modeling for MOSFETs and its application on 2D material field-effect transistors

Artificial Neural Network Compact Modeling Methodology for Complementary Field Effect Transistor

Compact modeling of metal–oxide TFTs based on the Bayesian search-based artificial neural network and genetic algorithm

Artificial Neural Network Based CNTFETs Modeling

Prediction of electrical properties of GAAFET based on integrated learning model

Modeling nanoscale MOSFETs by a neural network approach

Artificial neural network design for compact modeling of generic transistors

Physics-integrated Neural Network for Quantum Transport Prediction of Field-effect Transistor

A computationally efficient compact model for ferroelectric FETs for the simulation of online training of neural networks

A Combined N/PFET CFET-Based Design and Logic Technology Framework for CMOS Applications

AppGNN: Approximation-Aware Functional Reverse Engineering using Graph Neural Networks