On Robustness and Generalization of ML-Based Congestion Predictors to Valid and Imperceptible Perturbations

arXiv:2403.00103v1 Announce Type: new
Abstract: There is substantial interest in the use of machine learning (ML)-based techniques throughout the electronic computer-aided design (CAD) flow, particularly methods based on deep learning. However, while deep learning methods have achieved state-of-the-art performance in several applications, recent work has demonstrated that neural networks are generally vulnerable to small, carefully chosen perturbations of their input (e.g. a single pixel change in an image). In this work, we investigate robustness in the context of ML-based EDA …

abstract applications art arxiv cad computer congestion cs.ar cs.lg deep learning design electronic flow machine machine learning performance robustness state type work