Weighted Separator Function To Syndrome Depth¶
Claim/Theorem¶
This node is a source-grounded corollary schema rather than a named theorem from one paper. Let \(G_{\mathrm{hw}}\) be an \(N\)-vertex hardware graph of maximum degree at most \(\Delta_{\mathrm{hw}}\). Assume the hardware family admits a weighted separator function \(s_{\mathrm{hw}}(N)\) in the following sense: for every induced subgraph \(H\) of \(G_{\mathrm{hw}}\) with at most \(N\) vertices and every nonnegative vertex-weight function \(w\) on \(H\), there exists a separation \((A,B)\) of \(H\) with
and
Now suppose a local Clifford stabilizer-measurement circuit on \(G_{\mathrm{hw}}\) measures a bounded-weight local-expander quantum LDPC code with \(n\) data qubits. Put weight \(1\) on data-qubit vertices and weight \(0\) on ancillas. Then, provided \(s_{\mathrm{hw}}(N)=o(n)\), there exists a subset \(L\) of circuit qubits with \(|D\cap L|=\Theta(n)\) and
Applying [[expansion-cut-to-syndrome-depth.md]] yields
Therefore any sublinear weighted-separator profile on the hardware graph family forces superconstant syndrome-extraction depth for local-expander QLDPC families, and the static-grid law Omega(n/sqrt(N)) is just the specialization \(s_{\mathrm{hw}}(N)=Theta(\sqrt{N})\).
Dependencies¶
- [[expansion-cut-to-syndrome-depth.md]]
Conflicts/Gaps¶
- This is still a theorem about local Clifford stabilizer-measurement circuits, not arbitrary compilation maps or the full
CD(T_n,\mathfrak G)conjecture. - The argument requires bounded hardware degree to convert separator size into edge-boundary capacity.
- The balanced-cut conclusion is useful only when \(s_{\mathrm{hw}}(N)=o(n)\); otherwise the separator can absorb too many data qubits.
- The node is a meta-corollary: to use it one still needs a genuine weighted separator theorem for the chosen hardware family.
Sources¶
10.48550/arXiv.2109.1459910.1145/100216.100254