Track the frontier, from condensed matter to cosmology.
Search papers, labs, and topics across Lattice.
This paper introduces a novel quantum circuit synthesis technique that leverages mid-circuit measurements and feedforward to achieve exact $n$-qubit unitary implementation in constant quantum depth, a significant reduction from the prior $Θ(n)$ bound. This method utilizes single-qubit measurements, parity feedforward, and a deterministic construction with $O(n4^n\log(n+2))$ total qubits, enabling the exact synthesis of arbitrary unitaries. The key result demonstrates that by teleporting and coherently interfering the approximation error, constant-depth exact synthesis is achievable, offering a powerful new paradigm for quantum computation.
Exact quantum unitary synthesis is now possible in constant depth, slashing the required circuit depth from linear to $O(1)$ by cleverly using mid-circuit measurements and feedforward.
Mid-circuit measurements with feedforward reduce the optimal worst-case quantum depth for exact $n$-qubit unitary synthesis from $Θ(n)$ to $O(1)$, using arbitrary one- and two-qubit gates and unrestricted ancillary space. Measurements and resets without feedforward remain subject to the circuit causal-cone bound. Our deterministic construction uses single-qubit measurements, $O(1)$ rounds of parity feedforward, and $O(n4^n\log(n+2))$ total qubits, with all ancillas starting and ending in $|0\rangle$. Classical processing is not counted in quantum depth. We teleport only the matrix error of an approximation by Nehoran and Yuen. An exactly prepared ancillary state encodes this error, whose action is added back by coherent interference. At operator-norm error $2^{-n}$, one step of amplitude amplification suffices. Coherent fanout circuits achieve the same constant-depth bound, while circuits using only one- and two-qubit gates attain depth $Θ(n)$ with $O(4^n\log^2(n+2))$ gates. Generic targets require $Ω(4^n/(h+1))$ qubits at depth $h$, even under unrestricted classical control. Our depth-width trade-offs meet this bound within a factor $O(n^2)$ for $1\le h\le2^n$.