Neural Belief-Propagation Decoders for Quantum Error-Correcting Codes
Institut quantique, Université de Sherbrooke
Abstract: In this talk, I will introduce the decoding problem for error-correcting codes and its relationship to statistical inference and statistical physics. Then I will introduce the classical belief-propagation (BP) decoder, which works very well for classical codes but fails for quantum codes. Finally, I will show that the BP decoder can be trained as a neural network, leading to a substantial improvement for the quantum case. BP-based algorithms are widely used in many research areas, and our results imply that they can be adapted into the quantum world via machine-learning techniques.
Contact: Lei Wang 9853