DeepGate: Learning Neural Representations of Logic Gates*
TimeWednesday, July 13th2:15pm - 2:37pm PDT
Location3000, Level 3
Event Type
Research Manuscript
ML Algorithms and Applications
DescriptionThe success of Deep Learning paved its path to EDA, where many DL solutions are proposed to solve specific problems. However, \textit{How to learn a good circuit representation?} is still an open question. In this work, we propose DeepGate, a neural representation learner for logic gates. It learns an effective representation by exploiting the various inductive biases from circuit structure as domain knowledge and uses logic simulated probabilities as a rich supervision for every node. The Experimental results show that DeepGate learns a reasonable representation from probability prediction and can be applied to downstream tasks like test point insertion problem.