Problem
Research questions and friction points this paper is trying to address.
Hardware Trojan
Scalability
System-on-Chip
Gate-level netlist
Adversarial detection
ADVERSARIAL: And-Inverter Graph-Assisted Hardware Trojan Detection At Scale
🤖 AI Summary
This work addresses the scalability challenge in hardware Trojan detection for ultra-large-scale SoCs by proposing an efficient method that integrates And-Inverter Graphs (AIGs) with knowledge graph embedding (KGE). The approach models gate-level netlists as AIGs and leverages symbolic learning to exploit the bounded fan-in and uniform semantics of AIG structures, generating node representations that preserve multi-hop structural context. This enables linearly scalable detection complexity. By uniquely combining symbolic learning with AIG topology, the method effectively identifies rare, functionally anomalous connections as Trojan indicators. Experimental results on large-scale SoC benchmarks demonstrate a clear geometric separation between Trojan and benign nodes in the embedding space, validating both the scalability and practical utility of the proposed technique.
📝 Abstract
Modern System-on-Chip (SoCs) often contain hundreds of millions to tens of billions of gates, making existing Hardware Trojan (HT) detection methods impractical due to their immense scale. The proposed approach incorporates symbolically enabled learning by modeling flattened gate-level netlists as Boolean networks represented as And-Inverter Graphs (AIGs), where all internal nodes are 2-input AND gates and inversions reside on the edges. Each directed connection is expressed as a triple within a Knowledge Graph Embedding (KGE) framework, producing compact, constant-size per-node representations that retain multi-hop structural context. The AIG's bounded fan-in and uniform semantics ensure training and inference complexity scale linearly with edge count, addressing major scalability bottlenecks in HT detection. Symbolically enabled learning across deep datapaths enables the model to differentiate circuit structures from rare and functionally inconsistent connections that signify potential Trojan triggers and payloads. Experiments on large-scale SoC benchmarks demonstrate clear geometric separation between Trojan and benign nodes and practical scalability.
Innovation
Methods, ideas, or system contributions that make the work stand out.
And-Inverter Graph
Hardware Trojan Detection
Knowledge Graph Embedding
Scalable Boolean Networks
Symbolically Enabled Learning
🔎 Similar Papers
No similar papers found.
💼 Related Jobs
