Semi Supervised Learning of Multi-Dimensional Analog Circuits

Semi Supervised Learning of Multi-Dimensional Analog Circuits
Title:
Semi Supervised Learning of Multi-Dimensional Analog Circuits
Other Titles:
Design Automation Conference (DAC) (2019)
DOI:
Publication URL:
Publication Date:
02 June 2019
Citation:
Abstract:
Analog circuits are strictly designed under operational, functional and technology constraints. Together, these bounds creates a sparse multi-dimensional design optimization space with scarcity of labeled analog training data making supervised learning methods ineffective. Accurate approximation of multi-target analog circuits therefore requires generation of labeled data around dominant bias and with relevant variance. With such approach, we explore state-of-the-art semi supervised, generative adversarial network (GAN) towards analog performance modeling. We report on various multi-target analog circuit classification experiments and demonstrate stable GAN performance achieving 2-5% higher accuracy and utilizing only 10% fully simulated manually annotated labeled data against supervised learning methods.
License type:
PublisherCopyrights
Funding Info:
Description:
ISBN:

Files uploaded:
File Size Format Action
There are no attached files.