Neural Quantum States for Volume-Law Ground States

Accompanying code for "Comment on Can Neural Quantum States Learn Volume-Law Ground States?" (https://arxiv.org/abs/2309.11534).

Content of the repository

• dis_fermions.py: functions for generating the SYK-type disordered fermionic Hamiltonian (DF) and for diagonalizing it.

• intializers.py: initializer functions for generating random matrices.

• learning.py: function performing the training of the Neural Quantum State (NQS) ansätze.

• renyin.py: function computing the Rényi-n entanglement entropy.

• simple_model.py: flax model for the two-layer perceptron NQS (FF).

• sk.py: file containing functions for generating the quantum Sherrington-Kirkpatrick Hamiltonian (QSK) and for diagonalizing it.

• slater.py: flax model for the Slater determinant NQS with backflow transformation (FF+SD).

• runs/: folder containing the results of the simulations.

  • energy_params_disf_bf/, energy_params_disf_simple/, energy_params_sk/: folders containing the parameters of the optimized NQSs used to compute the energy errors for the two models.
  • df_entropy.out: Rényi-2 entropy of the exact DF ground state.
  • disf_bf.out: best infidelity for the optimization of FF+SD for the DF model.
  • disf_bf_energy_errors.py: computing the relative energy error of the optimized FF+SD for the DF model. Data in disf_bf_energy_errors.out.
  • disf_bf_runs.py: optimizing FF+SD for the DF model.
  • disf_simple.out: best infidelity for the optimization of FF for the DF model.
  • disf_simple_energy_errors.py: computing the relative energy error of the optimized FF for the DF model. Data in disf_simple_energy_errors.out.
  • disf_simple_runs.py: optimizing FF for the DF model.
  • figure.pdf: figure in the comment.
  • plot.py: plotting the data and creating the figure.
  • runs_entropy.py: computing the Rényi-2 entropy of the exact DF and QSK ground states.
  • sk_entropy.out: Rényi-2 entropy of the exact QSK ground state.
  • sk.out: best infidelity for the optimization of FF for the QSK model.
  • sk_energy_errors.py: computing the relative energy error of the optimized FF for the QSK model. Data in sk_energy_errors.out.
  • sk_runs.py: optimizing FF for the QSK model.