Automatic Differentiation for Complex Valued SVD
arXiv:1909.02659v317 citations
Originality Incremental advance
AI Analysis
This work provides a foundational component for automatic differentiation in complex-valued computations, which is incremental but essential for advancing tensor network methods.
The authors derived the backpropagation formula for complex-valued singular value decompositions (SVD), enabling automatic differentiation for complex numbers and applications in tensor networks.
In this note, we report the back propagation formula for complex valued singular value decompositions (SVD). This formula is an important ingredient for a complete automatic differentiation(AD) infrastructure in terms of complex numbers, and it is also the key to understand and utilize AD in tensor networks.