Physics Department - Data-Driven Advances in Quantum Error Correction and Control

Abstract
As quantum computing continues to face challenges from noise and errors, effective solutions are necessary for fault tolerance. This talk presents our data-driven approaches to quantum error correction (QEC) decoding and quantum control of many-body spin systems.
We start with QEC, showcasing how models trained on syndrome–error data reduce reliance on handcrafted priors. We explore predictive decoders mapping syndromes to corrections and generative decoders modeling error patterns, emphasizing their application in circuit-level simulations.
Next, we address quantum control, where traditional analytic methods become limiting as we approach performance thresholds. We introduce a simulation-guided stochastic tree search to navigate complex pulse-sequence spaces, revealing pulse sequences that outperform conventional designs in solid-state spin ensembles. This method expands our control options significantly, enabling us to tackle performance-limiting effects more effectively.
Together, these advancements demonstrate the power of data-driven methodologies in enhancing quantum error correction and control, paving the way for more robust quantum computing.