Abstract:

Standard fault-tolerant quantum error correction (QEC) schemes can protect quantum information against arbitrary errors on individual qubits, but with the implicit assumption that the underlying physical systems remain within the qubit subspace.  Fault-tolerance may therefore be lost under "leakage" errors (such as a photonic qubit being absorbed) which violate this assumption, and techniques for modifying QEC to cope with generic leakage can be very costly in terms of the extra operations required.  I will discuss a more specific "loss" model of leakage, applicable to many implementations (photonic ones especially), and show that it allows for fault-tolerance to be analysed and achieved at much less cost in common Calderbank-Shor-Steane (CSS) codes.

(PLEASE NOTE NON-STANDARD DATE)