Intrinsic Quantum Codes: One Code To Rule Them All
Eric Kubischta, Ian Teixeira
TL;DR
The paper reframes quantum error correction in intrinsic, representation-theoretic terms: a code is a subspace $\mathcal{I}$ of a representation $V$ of a symmetry group $G$, with errors organized by isotypic sectors of $\mathscr{L}(V)$. The central Schur-Bootstrap result shows that any $G$-equivariant embedding into a physical space $\mathcal{H}$ preserves the Knill-Laflamme conditions and yields a covariant extrinsic code $\mathcal{C}=\mathcal{U}(\mathcal{I})$ with the same error protection, plus transversality for the induced logical group $\mathsf{G}$. Defining intrinsic distance $d$ via the smallest tensor power in which irreps appear, the authors prove that a code of distance $d$ in the intrinsic setting yields a corresponding extrinsic code of distance $d$ under $G$-equivariant realizations. They illustrate this with a minimal SU(2) intrinsic code $\{\{ {\mathbf{5}^{\prime}} ,2,2\}\}$, which generates a family of extrinsic codes across qubit, qudit, bosonic, and rotor platforms all covariant under $\mathsf{G}=S_3$, and reveals moduli spaces like $\mathbb{C}P^{m-1}$ and a depth notion that refines conventional distance. This framework promises hardware-agnostic design of covariant quantum codes and systematic discovery of new codes across diverse physical implementations.
Abstract
Drawing on the analogy between intrinsic and extrinsic geometry, we define an intrinsic quantum code as a subspace of a group representation. We show how this abstract code rules over any physical (extrinsic) realization by dictating its error-correction properties. By finding one intrinsic code, we simultaneously uncover properties of all its realizations. This approach brings diverse codes into a unified framework and binds them through a single underlying symmetry.
