Recent progress in quantum error correction (QEC) and early fault-tolerant quantum computing (FTQC) has brought an imminent but fundamental question: whether errors in quantum operators should be fully corrected to achieve truly quantum utility. Achieving a fully error-corrected execution is not only costly in terms of the number of physical qubits with low error rate under the FTQC theorem, but may also not be necessary for many physics-informed workloads (e.g., quantum simulation of open quantum systems and efficient preparation of ground/thermal states). Open quantum systems evolve nonunitarily due to their interaction with the environment, creating a fundamental mismatch with the strictly unitary dynamics enforced by conventional fully corrected FTQC. Existing approaches need to further address this mismatch by embedding the system into larger Hilbert spaces or approximating nonunitary maps with unitary decompositions, which incur substantial overhead in qubit count, circuit depth, and control complexity.
We are rethinking this problem by challenging the assumption that all quantum errors must be uniformly eliminated; instead, we posit that for open quantum system simulations, a subset of noise processes can be selectively retained, shaped, and exploited to directly emulate the desired nonunitary dynamics. The team's recent work shows that, by leveraging intrinsic hardware noise, it is possible to obtain target non-unitary operators that bypass the need for large ancillary spaces and deep unitary decompositions. This shift reframes noise from a purely adversarial effect into a tunable computational resource, fundamentally altering the cost-accuracy tradeoff of open quantum system simulations.
However, leveraging the noise is application-specific; therefore, it calls for automated software that can realize noise-assisted quantum simulations with tunable fidelity for (partial) error-corrected logical quantum gates. Rather than enforcing a single, hardware-limited notion of "maximum fidelity", FTQTF automatically determines application-specific error-rate requirements and synthesizes logical operations whose residual noise channels are matched to the target open-system dynamics. The framework integrates (i) a system-dependent resource estimation tool (RET) to quantify which error components must be corrected versus retained, (ii) a logical-gate synthesis tool (LST) to realize continuously tunable effective noise channels, and (iii) a fidelity-bounded compilation tool (CCT) to ensure reliable execution under realistic noise drift. Together, these capabilities enable scalable, noise-aware simulation of open quantum systems with orders-of-magnitude lower overhead than fully corrected FTQC, while maintaining scientific accuracy.
This project will deliver an open-source software framework that directly supports the DoE mission in quantum simulation, scientific computing, and co-designed software ecosystems. By reducing the cost of QEC, enabling precision-aware and noise-adaptive execution, and providing rigorous end-to-end fidelity guarantees, FTQTF establishes a scalable pathway for simulating molecular, materials, and device dynamics under realistic environmental conditions.
