AI Paper Atlas

Research map / Graph learning / Graph algorithms and clustering

A Hierarchical Circuit Symbolic Discovery Framework for Efficient Logic Optimization

Yinqi Bai, Jie Wang, Tong Xialiang, Longdi Pan, Jilai Pan, Mingxuan Yuan, Jianye HAO

ICLR 2026 · Graph learning › Graph algorithms and clustering · approach: Neural combinatorial optimization on graphs

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add hierarchical layer-by-layer symbolic tree learning (ablation)HIS · six circuits (avg recall)prediction recall: Hyp 0.81 -> 0.82; Square 0.74 -> 0.94; DesPerf 0.77 -> 0.83; Ethernet 0.81 -> 0.99Improved
add group reward advantage in PPO (ablation)HIS · six circuitsprediction recall: DesPerf 0.51 -> 0.83; Hyp 0.88 -> 0.82Mixed
add tree-structured parent and sibling aggregation (ablation)HIS · six circuitsprediction recall: Multiplier 0.51 -> 0.94; Ethernet 0.91 -> 0.99Improved
use HIS scoring function in Mfs2 with k=50%Default Mfs2 · six challenging circuitsruntime: average runtime reduction of 40.27%Improved
use HIS scoring function in Mfs2 with k=50%Default Mfs2 · Hypruntime (s): 265.93 -> 85.46Improved
apply HIS-Mfs2 twice (2HIS) with k=40%HIS-Mfs2 · six circuitsoptimized size and depth: average reduction of 7.43% in size and depth; runtime reduced by 7.82%Improved
apply HIS-Mfs2 twice (2HIS) with k=30%HIS-Mfs2 · six circuitsoptimized size and runtime: average reduction of 6.95% in size and depth; runtime reduction of 27.22%Improved
replace COG GNN scorer with HIS symbolic scorerCOG-Mfs2 · six circuitsonline runtime: 22.91% average runtime improvementImproved
replace CMO scorer with HISCMO-Mfs2 · six circuitsonline runtime: 11.96% average runtime improvementImproved
replace GNN scorer COG with HIS symbolic scorerCOG · EPFL circuitsinference time: 296x speedupImproved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy