A regularized constrained Newton optimizer on the sphere of normalized tensors reaches lower negative log-likelihood faster than gradient descent for tensor-network Born machines.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Regularized second-order optimization of tensor-network Born machines
A regularized constrained Newton optimizer on the sphere of normalized tensors reaches lower negative log-likelihood faster than gradient descent for tensor-network Born machines.