# US 2026/0080295 A1 · Methods and systems for error correction in neutral atom quantum computers

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2026/0080295 A1
- Document ID: US20260080295A1
- Status: Published application
- Kind: A1
- Application number: 19/093,045
- Filed: 2025-03-27
- Published: 2026-03-19
- Earliest claimed: 2023-03-03
- Assignees: Atom Computing Inc.
- Applicants: Atom Computing Inc.
- Inventors: Daniel CROW
- Web page: https://patentgrep.com/patent/US20260080295A1

## CPC

- G06N 10/70: Quantum error correction, detection or prevention, e.g. surface codes or magic state distillation
- G06N 10/20: Models of quantum computing, e.g. quantum circuits or universal quantum computers
- G06N 10/40: Physical realisations or architectures of quantum processors or components for manipulating qubits, e.g. qubit coupling or qubit control
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

A method for error corrected quantum computation may include identifying that a qubit has been lost; replacing the qubit; reimplementing the qubit into the circuit; and flagging measurements taken while the qubit was missing as untrustworthy.

## Claims

### Claim at position 1

-122. (canceled)

### Claim 123

A method for error corrected quantum computation, the method comprising:
(a) identifying a lost qubit within an array of qubits;
(b) replacing said lost qubit with a replacement qubit;
(c) reimplementing said replacement qubit into a quantum circuit; and
(d) flagging measurements taken while said lost qubit was missing as untrustworthy.

### Claim 124 (depends on 123)

The method of claim 123, wherein said identifying in (a) comprises using a plurality of swap gates.

### Claim 125 (depends on 124)

The method of claim 124, wherein a swap gate within said plurality of swap gates is implemented as a plurality of CNOT gates.

### Claim 126 (depends on 124)

The method of claim 124, further comprising:
measuring alternating qubits in a lattice;
performing said plurality of swap gates to transfer data stored on data qubits to ancilla qubits; and
measuring said data qubits to identify said lost qubit in (a).

### Claim 127 (depends on 123)

The method of claim 123, wherein identifying in (a) comprises using a modified knock-knock protocol, wherein said modified knock-knock protocol comprises:
providing a first qubit to be probed using a second qubit, wherein said second qubit is an ancilla qubit;
preparing said second qubit in a |+>state;
applying a control-Z gate between said first qubit and said second qubit;
rotating said second qubit back to a computational basis; and
performing a measurement on said second qubit to identify said lost qubit in (a).

### Claim 128 (depends on 123)

The method of claim 123, wherein said reimplementing in (c) comprises: (i) at a decoder algorithm, taking in a graph and determining a set of edges, a set of hyperedges, or both.

### Claim 129 (depends on 128)

The method of claim 128, wherein, prior to (i), the method comprises updating a matching graph passed to said decoder algorithm based at least in part on a predicted probability distribution of a lost qubit replaced in (b).

### Claim 130 (depends on 129)

The method of claim 129, wherein said decoder algorithm is a minimum-weight perfect matching decoder algorithm, union find, tensor network decoder, belief propagation with ordered statistics decoder, maximum likelihood decoder, or a look up table decoder.

### Claim 131 (depends on 130)

The method of claim 130, wherein said decoder algorithm is a minimum-weight perfect matching decoder algorithm, and wherein the method further comprises: updating a matching graph passed to said minimum-weight perfect matching decoder algorithm based on a predicted probability distribution of said lost qubit.

### Claim 132 (depends on 131)

The method of claim 131, further comprising:
if an ancilla qubit is lost, updating said matching graph so that a node involving said ancilla qubit is connected by edges corresponding to said predicted probability distribution of said lost qubit; and
if a data qubit is lost, updating said matching graph by assigning said predicted probability distribution to each node involving said data qubit.

### Claim 133 (depends on 123)

The method of claim 123, further comprising performing (a)-(d) during a quantum computation circuit.

### Claim 134 (depends on 123)

The method of claim 123, wherein (d) comprises flagging a measurement taken during a window of time that includes a time when said lost qubit was missing as untrustworthy.

### Claim 135 (depends on 123)

The method of claim 123, wherein said lost qubit, said replacement qubit, or both is a trapped atom qubit.

### Claim 136 (depends on 135)

The method of claim 135, wherein said trapped atom qubit is a neutral atom qubit, and wherein said trapped atom qubit comprises qubit states comprising nuclear spin states.

### Claim 137 (depends on 136)

The method of claim 136, wherein said neutral atom qubit is a Group II element or a Group II-like element.

### Claim 138 (depends on 123)

The method of claim 123, wherein said identifying in (a) comprises an operation in which a two-qubit interaction between a qubit and said lost qubit has an effect of a Pauli operation or an identity operation on said qubit.

### Claim 139 (depends on 138)

The method of claim 138, wherein said two-qubit interaction comprises an excitation of a nuclear spin state of a neutral atom to a Rydberg state of said neutral atom.

### Claim 140 (depends on 123)

The method of claim 123, wherein (a)-(d) comprise a portion of an error correcting code, and wherein said error correcting code comprises a topological code, a stabilizer code, a surface code, a color code, a toric code, a shor style code, or a qLDPC code.

### Claim 141 (depends on 123)

The method of claim 123, wherein said reimplementing at (c) comprises: implementing an error correction code, wherein an implementation of said error correcting code comprises a decoder, wherein said decoder is configured to receive a matching graph and to determine a set of edges, and wherein said matching graph received by said decoder is updated based on a predicted probability distribution of said lost qubit.

### Claim 142 (depends on 141)

The method of claim 141, wherein each node in said matching graph corresponds to a change-of-value of a particular stabilizer, and wherein pairs of nodes are connected by edges corresponding to possible physical errors.

### Claim 143 (depends on 142)

The method of claim 142, wherein said edges are weighted based on a likelihood of a particular error occurring.
