Unleashing Optimizations in Dynamic Circuits through Branch Expansion

Dynamic quantum circuits enable adaptive operations through intermediate measurements and classical feedback. Current transpilation toolchains, such as Qiskit and T$\ket{\text{ket}}$, however, fail to fully exploit branch-specific simplifications. In this work, we propose recursive branch expansion as a novel technique which systematically expands and refines conditional branches. Our method complements existing transpilers by creating additional opportunities for branch-specific simplifications without altering the overall circuit functionality. Using randomly generated circuits with varying patterns and scales, we demonstrate that our method consistently reduces the depth and gate count of execution paths of dynamic circuits. We also showcase the potential of our method to enable optimizations on error-corrected circuits.