Conceptual
Browse Instruct Curate

Transfer Learning Analysis of Variational Quantum Circuits

A formalism for analyzing transfer learning in variational quantum circuits: the adaptation of a VQC pretrained in one domain to a new domain is expressed as the transition of its 1-parameter unitary subgroups exp(i*theta*H), and loss bounds characterize how adaptable and capable the transferred circuit is. The theory gives a heuristic account of how knowledge transfers between domains and yields an analytical fine-tuning method computing the optimal parameter transition for adaptation between similar domains.