Saved in:
Bibliographic Details
Main Authors: Lee, Sungyoung, Wang, Ziyi, Kim, Seunggeun, Lee, Taekyun, Lai, Yao, Pan, David Z.
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2502.08949
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913846507077632
author Lee, Sungyoung
Wang, Ziyi
Kim, Seunggeun
Lee, Taekyun
Lai, Yao
Pan, David Z.
author_facet Lee, Sungyoung
Wang, Ziyi
Kim, Seunggeun
Lee, Taekyun
Lai, Yao
Pan, David Z.
contents Pretraining models with unsupervised graph representation learning has led to significant advancements in domains such as social network analysis, molecular design, and electronic design automation (EDA). However, prior work in EDA has mainly focused on pretraining models for digital circuits, overlooking analog and mixed-signal circuits. To bridge this gap, we introduce DICE, a Device-level Integrated Circuits Encoder, which is the first graph neural network (GNN) pretrained via self-supervised learning specifically tailored for graph-level prediction tasks in both analog and digital circuits. DICE adopts a simulation-free pretraining approach based on graph contrastive learning, leveraging two novel graph augmentation techniques. Experimental results demonstrate substantial performance improvements across three downstream tasks, highlighting the effectiveness of DICE for both analog and digital circuits. The code is available at github.com/brianlsy98/DICE.
format Preprint
id arxiv_https___arxiv_org_abs_2502_08949
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle DICE: Device-level Integrated Circuits Encoder with Graph Contrastive Pretraining
Lee, Sungyoung
Wang, Ziyi
Kim, Seunggeun
Lee, Taekyun
Lai, Yao
Pan, David Z.
Machine Learning
Pretraining models with unsupervised graph representation learning has led to significant advancements in domains such as social network analysis, molecular design, and electronic design automation (EDA). However, prior work in EDA has mainly focused on pretraining models for digital circuits, overlooking analog and mixed-signal circuits. To bridge this gap, we introduce DICE, a Device-level Integrated Circuits Encoder, which is the first graph neural network (GNN) pretrained via self-supervised learning specifically tailored for graph-level prediction tasks in both analog and digital circuits. DICE adopts a simulation-free pretraining approach based on graph contrastive learning, leveraging two novel graph augmentation techniques. Experimental results demonstrate substantial performance improvements across three downstream tasks, highlighting the effectiveness of DICE for both analog and digital circuits. The code is available at github.com/brianlsy98/DICE.
title DICE: Device-level Integrated Circuits Encoder with Graph Contrastive Pretraining
topic Machine Learning
url https://arxiv.org/abs/2502.08949