⌅
H|1⇤ = (|0⇤ |1⇤)/ 2 ⇥ | ⇤
$#.%"',,*,"+*,,'+%.*)"+*3'7
Overview: Quantum Error Correction Lectures
4#.%5-;.:"+*3'7
1'+*)35;'=';"')+*3.)>
1 - Introduction to basic (non-topological)
quantum error correcting (QEC) codes
• General challenges of QEC
• 3-qubit code, concatenation
• 9-qubit Shor code
• Stabiliser QEC codes
• Elements of quantum fault-tolerance
2 - Introduction to topological QEC codes
• Kitaev’s toric code
•Surface code 7
""4:2(0'5-;.:"+*3'7
6$#.%":2(0'"()3"#.%5-;.:"+*3'
• Logical gate operations<.,0%5;'=';"')+*3.)>
3":2(0'5-;.:"*)"()/"6$#.%
3 - Scalable topological architectures
!
• Coupling logical qubits in 2D layouts
Figure 5. Arrangements of physical qubits for rough lattice merging. Left and
• Lattice surgery-based CNOT gate right continuous surfaces encode separate logical qubits. The pink qubits form
the intermediate qubit line for the merging operation.
Why Quantum Error Correction?
Motivation: Execution of quantum algorithms to solve non-trivial problems requires
control over large registers of qubits, and the application of large number of quantum gates
Grover’s search
Shor’s factoring algorithm
algorithm
Why Quantum Error Correction?
Problems: Qubits of the quantum computer cannot be perfectly isolated from the
environment, control fields can’t be implemented perfectly…
Thus, the operations of the quantum hardware can be faulty.
This leads to errors in quantum gates, initial state preparation, measurements.
Error sources
These errors rapidly
|0
propagate and |0
accumulate. |0
This inhibits the reliable
execution of long
computations on
large qubit registers |0
Classical error correction
protection of classical ...011010...
information by redundancy
...011010... ...011010... ...011010...
Example: n-bit repetition code, with majority vote as correction strategy
0 —> 000 001 0 —> 00000 00110
1 —> 111 110 1 —> 11111 11001
The repetition code fails if more than half of the bits suffer errors (bit flips)
Generalisation to the quantum case is challenging…
Can’t copy quantum states for redundancy:
forbidden by the no-cloning theorem for quantum states
Generalisation to the quantum case is challenging…
Can’t directly measure quantum states to find out about the errors
that occurred. This leads to a collapse of the wave function
↵|0i + |1i |0i or |1i …information about
↵ and lost
measure (Z-basis)
even worse for entangled states
↵|000i + |111i |000i or |111i …by measuring one
of the qubits only
Generalisation to the quantum case is challenging…
Errors are more diverse in the quantum world: Error operator
|0i |1i bit flip errors (like for classical bits)
But there are also phase flip errors
↵|0i + |1i ! ↵|0i |1i
or both types of errors simultaneously
↵|0i + |1i ! ↵|1i |0i
Generalisation to the quantum case is challenging…
Errors in quantum information are intrinsically continuous (rather than discrete):
Example: Undesired residual perturbation Hamiltonian H/X
i X
U (t) = exp( iHt) = e
Effect on an initial state | i
Probability of bit-flip error
Probability of no error
Effective error model we use in the following:
Stochastic, independently distributed X, Y, Z errors with probability p
Conceptual setting of a quantum error correcting code
) *+
'
,+
) *+
| i
|0
'
Correction
Encoding
,+ Quantum state
|0
recovered ?
Errors
...
7
8
|0
)-/
8
7
8
)-/
8
A medical analogy
| iL | iL
| i
|0
Error syndrome measurement
Encoding
|0
Correction
Errors
...
|0
Disease
|0 M
...
|0 M
From the syndrome:
Which errors have happened? Which cure to give?