We introduce a neural network impurity solver for real-frequency dynamical mean-field theory that employs a multihead cross-attention mechanism to map hybridization functions to spectral functions, conditioned on impurity parameters. Trained on high-quality matrix product states data from complex contour time evolution and incorporating derivative constraints with respect to the complex-time angle, our model achieves smooth generalization to the real-frequency axis. Benchmarking on the single-band Hubbard model for the Bethe lattice demonstrates quantitative accuracy across metallic, strongly correlated, and insulating regimes.

Close

PrevNext

View figure in article

• Electronic structure
• Strongly correlated systems
• Dynamical mean field theory
• Machine learning