Last update: August 2026
Errors are a fundamental challenge in quantum computing. As systems grow and computations become deeper, reliable execution requires quantum information to be encoded, errors to be detected and corrected, and logical operations to be performed while preserving this protection.
In my group, we investigate quantum error correction and fault-tolerant computation, with a focus on new code constructions, practical decoding, and the architectural requirements of reliable quantum computing. Our work spans the design and analysis of quantum codes, efficient syndrome processing, and methods for performing computations on encoded quantum information.
Quantum error correction protects logical quantum information by distributing it across multiple physical qubits. Useful codes need strong error-correcting properties, but their structure also determines how efficiently they can be encoded, decoded, and used for computation.
One direction of our work investigates holographic quantum error-correcting codes, based on tensor-network structures originating from quantum information and holography. Starting from hyperinvariant tensor networks, we have developed and analyzed new code constructions and studied their encoding rate, distance, decoding properties, and logical operations.
We also investigate how these codes behave under different physical noise models. This includes biased noise, where some types of errors occur more frequently than others, as is common in realistic quantum hardware. These studies help us understand which structural properties of a code are relevant under realistic operating conditions.
A quantum error-correcting code also requires an efficient decoder.
During computation, syndrome measurements provide information about errors in the encoded state. The decoder processes this information and determines an appropriate correction. For large fault-tolerant systems, this classical computation has strict requirements on latency, throughput, memory, and implementation complexity.
We develop scalable decoding methods with these constraints in mind. One direction explores hyperdimensional computing for real-time decoding and studies how classical processing can be integrated into the control architecture of a quantum computer.
Our work also considers the circuits used for syndrome extraction. Reducing their depth and complexity can lower the overhead of error correction and simplify practical implementations.
Protecting quantum information is only useful if computations can also be performed on the encoded state.
We investigate fault-tolerant logical operations and code constructions that support them efficiently. In particular, we study how holographic and heterogeneous code structures can provide logical operations while maintaining error protection, including approaches toward universal fault-tolerant computation.
This connects code design directly to architecture and compilation. Error correction determines qubit and communication requirements, decoding creates a substantial classical computing workload, and logical operations introduce their own constraints on execution.
Our aim is to understand these interactions and develop reliability as an integral part of full-stack quantum system design.
Biased-noise thresholds of zero-rate holographic codes with tensor-network decoding. Physical Review A, 2026.
Far from perfect: Quantum error correction with (hyperinvariant) Evenbly codes. Quantum, 2025.
A scalable real-time decoder for quantum error correction based on hyperdimensional computing. IEEE International Conference on Quantum Computing and Engineering (QCE), 2025.
Universal fault-tolerant logic with heterogeneous holographic codes. Preprint, 2025.
Holographic codes from hyperinvariant tensor networks. Nature Communications, 2023.
Low-depth flag-style syndrome extraction for small quantum error-correction codes. IEEE International Conference on Quantum Computing and Engineering (QCE), 2023.
LEGO_HQEC: Software for constructing and studying holographic quantum error-correcting codes, including tensor-network and optimization-based decoding methods.
[GitHub] [Publication]