CircuitGate is proposed, a function-aware AIG representation learning framework that advances from gate-level semantics to circuit-level functional modeling, and achieves the best equivalent-gate identification performance, demonstrating strong cross-dataset generalization.
Abstract
And-Inverter Graphs (AIGs) are fundamental representations for logic synthesis and verification in Electronic Design Automation (EDA). As structured representations of complex digital systems, AIGs require models to capture functional dependencies beyond local structure and remain robust to functionality-preserving transformations. In learning-based AIG representation, existing approaches are predominantly based on GNNs and rely on local gate-level message passing, limiting their ability to capture circuit-level functional context and making the learned representations sensitive to topology-specific patterns. Therefore, we propose CircuitGate, a function-aware AIG representation learning framework that advances from gate-level semantics to circuit-level functional modeling. CircuitGate explicitly encodes global primary-input (PI) support and models support-overlap-aware reconvergence between fanins, while incorporating logic-inspired Boolean constraints to encourage functionally consistent representations. We evaluate CircuitGate on the large-scale ForgeEDA benchmark and further validate it on the EPFL and ITC'99 benchmarks. Across equivalent-gate identification and signal-probability prediction tasks, CircuitGate consistently outperforms existing methods, achieving up to 21.7% and 14.2% reductions in MAE, respectively. Under direct ForgeEDA-to-OpenABC transfer without fine-tuning, CircuitGate also achieves the best equivalent-gate identification performance, demonstrating strong cross-dataset generalization. These results demonstrate the effectiveness of modeling circuit-level functional dependencies beyond local topology.
LevelSyn is a novel physical-aware logic synthesis framework that integrates hierarchical representation learning with a wirelength-driven optimization engine that significantly outperforms state-of-the-art methods on the EPFL benchmark suite.
Jing-Yi Zhou, Zheng-Yuan Shi, Zi-Yang Zheng et al.· 0 citations
DeepSeq3 is introduced, a novel hierarchical framework that abstracts circuits into a two-level representation: fine-grained combinational subgraphs partitioned by flip-flops (FFs) and a high-level Super-Node Graph (SNG) that models the register-transfer structure.
Jing-Yi Zhou, Zhengyuan Shi, Jiaying Zhu et al.· 0 citations
A benchmark built from human-designed, open-source circuits associated with Tiny Tapeout manufacturing shuttles is introduced, exposing distinct visual and design capabilities and, in one family, a policy rather than capability limit.
E-ALS is proposed, an e-graph-based framework for approximation-aware structural search that combines Function-Reduced Saturation, an ALS-coupled surrogate, search-based extraction, and budget-guided refinement to identify approximation-friendly equivalent structures.
Bin Sun, Jia-Nan Mu, Jia-Xi Zhang et al.· 0 citations
QaiJi IR, an eight-layer family organized by a five-axis semantic contract, is presented, an eight-layer family organized by a \emph{five-axis semantic contract} for data, semantics, lowering, runtime, and verification.
Following Moore's Law, the ever-increasing complexity, i.e., the number of logic gates, of digital systems poses significant challenges for verification at the netlist (i.e., logic gate) level. On the one hand, the inherent bit-level parallelism exposed by netlists cannot be efficiently exploited to accelerate simulati...
M. Weißbrich, Mohamed Wassim Chebili, G. Payá-Vayá· IEEE International Conferenc...· 0 citations
Related blog posts
MIT News · Artificial Intelligence· news.mit.eduNov 17, 2026
By studying the computational foundations of language learning, Associate Professor Jacob Andreas builds intelligent systems that can communicate effectively with humans.
Microsoft Research Blog· microsoft.comNov 16, 2026
AI coding agents are increasingly help developers fix bugs and build features largely on their own, but every interaction with their models consumes tokens. In each session, an agent sends instructions, tool descriptions, and session history to its model. Most efforts to reduce this cost focus on individual tasks. Yet developers return to the same […] The post TokenEconomizer: Learning from Experience to Make AI Coding Agents More Efficient appeared first on Microsoft Research.
MIT News · Artificial Intelligence· news.mit.eduNov 13, 2026