BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:GNN-ACLP: Graph Neural Networks Based Analog Circuit Link Predicti
 on - Guanyuan Peter Pan
DTSTART:20250821T114500Z
DTEND:20250821T124500Z
UID:TALK234820@talks.cam.ac.uk
CONTACT:Pietro Lio
DESCRIPTION:Circuit link prediction identifying missing component connecti
 ons from incomplete netlists is crucial in analog circuit design automatio
 n. However\, existing methods face three main challenges: 1) Insufficient 
 use of topological patterns in circuit graphs reduces prediction accuracy\
 ; 2) Data scarcity due to the complexity of annotations hinders model gene
 ralization\; 3) Limited adaptability to various netlist formats. We propos
 e GNN-ACLP\, a graph neural networks (GNNs) based method featuring three i
 nnovations to tackle these challenges. First\, we introduce the SEAL (lear
 ning from Subgraphs\, Embeddings\, and Attributes for Link prediction) fra
 mework and achieve port-level accuracy in circuit link prediction. Second\
 , we propose Netlist Babel Fish\, a netlist format conversion tool leverag
 ing retrieval-augmented generation (RAG) with a large language model (LLM)
  to improve the compatibility of netlist formats. Finally\, we construct S
 piceNetlist\, a comprehensive dataset that contains 775 annotated circuits
  across 10 different component classes. Experiments demonstrate accuracy i
 mprovements of 16.08% on SpiceNetlist\, 11.38% on Image2Net\, and 16.01% o
 n Masala-CHAI compared to the baseline in intra-dataset evaluation\, while
  maintaining accuracy from 92.05% to 99.07% in cross-dataset evaluation\, 
 exhibiting robust feature transfer capabilities.
LOCATION:Zoom (Meeting ID: 873 5011 9733\, Passcode: 732177)
END:VEVENT
END:VCALENDAR