Alignment Unlocks Complementarity: A Framework for Multiview Circuit Representation Learning
This addresses a critical barrier in multiview circuit representation learning for AI and hardware design, offering a novel solution to fuse complementary structural information, though it is incremental in improving self-supervised techniques.
The paper tackles the problem of multiview learning on Boolean circuits, where structural heterogeneity between views like AIG and XMG hinders effective fusion in self-supervised methods like masked modeling, by introducing MixGate, a framework that uses functional alignment as a precondition to enable multiview masked modeling, resulting in transformed performance from ineffective to powerful.
Multiview learning on Boolean circuits holds immense promise, as different graph-based representations offer complementary structural and semantic information. However, the vast structural heterogeneity between views, such as an And-Inverter Graph (AIG) versus an XOR-Majority Graph (XMG), poses a critical barrier to effective fusion, especially for self-supervised techniques like masked modeling. Naively applying such methods fails, as the cross-view context is perceived as noise. Our key insight is that functional alignment is a necessary precondition to unlock the power of multiview self-supervision. We introduce MixGate, a framework built on a principled training curriculum that first teaches the model a shared, function-aware representation space via an Equivalence Alignment Loss. Only then do we introduce a multiview masked modeling objective, which can now leverage the aligned views as a rich, complementary signal. Extensive experiments, including a crucial ablation study, demonstrate that our alignment-first strategy transforms masked modeling from an ineffective technique into a powerful performance driver.