Finally, this was a fantastic collaboration with Yingjia Lin, who led the project, and Ken, whose insights were very helpful throughout.

Excited to share this, and we welcome any comments!

We believe that this idea can be extended beyond the toric code to other code families. Exploring these directions may lead to new pathways for reducing measurement overhead in large-scale fault-tolerant quantum computation.

Through numerical simulation under circuit-level noise models for the toric code, we show that our scheme can improve decoding performance compared to conventional checks when using sliding-window decoding with a reduced window size.

In this work, we introduce local single-shot checks by partitioning the toric code into small patches, avoiding high-weight checks. With a dynamic measurement scheme, we show that the number of required measurement rounds can be reduced.

Quantum error correction usually needs multiple rounds of syndrome extraction to overcome measurement noise. Single-shot error correction uses just one round, but most codes require high-weight checks, which significantly degrade, and often eliminate, single-shot performance at the circuit level.

Excited to share this with the community, and we welcome any feedback!

We believe this opens up new possibilities for hardware–software co-design on the path toward a fully scalable quantum computer.

This was a fantastic collaboration with Sahil, who drove the project, and Ken and Jonathan, whose insights were invaluable throughout.

Our proposal moves away from the existing 2D-grid design toward a circular topology. We show that using such an architecture enables significant spacetime savings when using high-rate codes such as HGP and BB, as well as a substantial reduction in logical error rates compared to the 2D grid.

Finally, this project was a very fun collaboration with Dmitrii, who is an amazing undergraduate student at Duke, and Ken, without whom this would not have been possible.

Our method represents a resource-efficient way to estimate observables and has applications in both fault-tolerant and near-term quantum computing. Please check it out (arxiv.org/abs/2505.11486) — we welcome any comments!

and use this to build unbiased estimators from weighted expectation values. We then analyze when this method works and validate it via simulations, including time evolution under the Ising Hamiltonian.

Unitary errors, e.g., from fault-tolerant compilation, systematically bias observable estimates. The usual way to correct this bias is to use more non-Clifford gates. Here, we present an alternative strategy: decompose the ideal channel into a probabilistic mixture of noisy channels