Decoding Quantum Low Density Parity Check Codes
summary
Quantum LDPC codes promise ~50 physical qubits per logical qubit but require general decoders because matching fails; plain belief propagation often produces trapping sets and no threshold, while BP+OSD restores a threshold and gives about five orders of magnitude lower logical error at 1% in examples. Circuit-level noise blows up decoding graphs (one case: 144 → 8785-column decoding matrix), and the OSD post‑processing is cubic and slow, so the open problem is a faster OSD alternative.
AI-written summary · about
chapters (11)
- 0:00Quantum LDPC codes intro
- 2:03Why use quantum LDPC codes
- 3:09Talk structure
- 11:29Classical error correction
- 15:05Quantum error correction
- 22:44Belief propagation limits
- 23:56BP plus OSD decoder
- 28:55Circuit noise challenges
- 33:09BP+OSD performance
- 34:12Open problems in decoding
- 48:07Q&A
related talks
32mTowards local testability for quantum coding
59mDawn in the age of quantum fault tolerance
03mQuantum Algorithms, Complexity, and Fault Tolerance
44mNLTS Hamiltonians (and Strongly-Explicit SoS Lower Bounds) from Low-Rate Quantum LDPC Codes
01h 03mQuantum LDPC Codes with Transversal CCZ Gates
01h 01mBuilding a Quantum Computer with QLDPC Codes
51mPeaked quantum advantage using error correction
35mHalving the cost of surface codes by running error correction at the logical level