GAINESIS: Generative Artificial Intelligence NEtlists SynthesIS

Abstract

A significant problem in the field of hardware security consists of hardware trojan (HT) viruses. The insertion of HTs into a circuit can be applied for each phase of the circuit chain of production. HTs degrade the infected circuit, destroy it or leak encrypted data. Nowadays, efforts are being made to address HTs through machine learning (ML) techniques, mainly for the gate-level netlist (GLN) phase, but there are some restrictions. Specifically, the number and variety of normal and infected circuits that exist through the free public libraries, such as Trust-HUB, are based on the few samples of benchmarks that have been created from circuits large in size. Thus, it is difficult, based on these data, to develop robust ML-based models against HTs. In this paper, we propose a new deep learning (DL) tool named Generative Artificial Intelligence Netlists SynthesIS (GAINESIS). GAINESIS is based on the Wasserstein Conditional Generative Adversarial Network (WCGAN) algorithm and area–power analysis features from the GLN phase and synthesizes new normal and infected circuit samples for this phase. Based on our GAINESIS tool, we synthesized new data sets, different in size, and developed and compared seven ML classifiers. The results demonstrate that our new generated data sets significantly enhance the performance of ML classifiers compared with the initial data set of Trust-HUB.

A Cost-Driven Method for Deep-Learning-Based Hardware Trojan Detection

“…For deep learning, the size and quality of the dataset are two important fundamental prerequisites. Most of the existing methods use open-source data, such as the RS232-series provided by Trust-Hub, and it is worth mentioning that Liakos et al [40] proposed a new tool, GAINESIS, a WCGAN-based algorithm that synthesizes new samples for experiments. This is a convenient tool for models where it is difficult to extract a sufficient number of samples.…”

Hardware Trojan Insertion Using Reinforcement Learning

“…Such detectors need significant data (various HTinfected and HT-free circuits) to provide fair and unbiased results. An important problem observed in this space is the imbalance between HT-infected circuits versus HT-free ones, which negatively impacts the quality of the training data [23]. Subsequently, trained detectors can be biased towards favoring circuits as HT-infected, as we will see later in Section VII-B.…”

“…Liakos et al [23] employed GANs (Generative Adversarial Networks) to replicate the characteristics of HT-free and HTinfected circuits characteristics, such as structural, power, and timing features. The study utilizes a feature generative approach with GANs to generate synthetic data for training more effective HT detectors.…”

A Non-Destructive Method for Hardware Trojan Detection Based on Radio Frequency Fingerprinting

“…Because of the threat of HTs, Trojan-detection methods are increasingly being emphasized. In the pre-silicon phase, static-detection techniques can extract and analyze the characteristics of the gate-level netlists to identify suspicious networks without logical or functional simulation, and machine learning methods can be used to classify the unknown networks into Trojan and normal networks efficiently [2][3][4][5]. However, these techniques make it difficult to build perfect models for all ICs.…”