Research map / Probabilistic and scientific modeling / Molecules and proteins
Systematic improvement of neural network quantum states using Lanczos
Hongwei Chen, Douglas Gerard Hendry, Phillip E Weinberg, Adrian Feiguin
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add one Lanczos step (p=1) to symmetrized RBM (ablation) | symmetrized RBM · J1-J2 10x10 J2=0.5 | energy per site: -0.49580(2) -> -0.4968(4) | Improved |
| add one Lanczos step (p=1) to symmetrized RBM (ablation) | symmetrized RBM · J1-J2 10x10 J2=0.55 | energy per site: -0.48410(3) -> -0.4859(5) | Improved |
| add one Lanczos step (p=1) to symmetrized RBM (ablation) | symmetrized RBM · J1-J2 10x10 J2=0.45 | energy per site: -0.50916(2) -> -0.5099(5) | Improved |
| add one Lanczos step (p=1) to symmetrized RBM (ablation) | symmetrized RBM · J1-J2 6x6 J2=0.5 | energy per site: -0.50364(2) -> -0.50376(3) | Improved |
| increase Lanczos steps from p=1 to p=2 (ablation) | symmetrized RBM · J1-J2 6x6 J2=0.5 | energy per site: -0.50376(3) -> -0.50378(4) | Improved |
| increase Lanczos steps from p=1 to p=2 (ablation) | symmetrized RBM · J1-J2 6x6 J2=0.6 | energy per site: -0.49313(5) -> -0.49318(5) | No clear effect |
| add Lanczos steps (p=1,2) to symmetrized RBM (ablation) | symmetrized RBM · J1-J2 6x6 J2=0.6 | energy per site: -0.49298(5) -> -0.49318(5) | Improved |
| use symmetrized RBM wave function instead of CNN | symmetrized RBM · J1-J2 6x6 J2=0.5 | energy per site: -0.50185(1) -> -0.50364(2) | Improved |
| use symmetrized RBM wave function instead of CNN | symmetrized RBM · J1-J2 10x10 J2=0.55 | energy per site: -0.48277(1) -> -0.48410(3) | Improved |
| use symmetrized RBM instead of CNN, relative error | symmetrized RBM · J1-J2 6x6 frustrated regime | relative energy error: 10^-3 -> 10^-4 | Improved |
| compare Lanczos RBM with RBM+PP | RBM + Lanczos · J1-J2 10x10 J2=0.5 | energy per site: -0.4968(4) vs -0.497629(1) | Hurt |
| compare symmetrized RBM with RBM+PP | symmetrized RBM · J1-J2 6x6 J2=0.55 | energy per site: -0.49501(1) vs -0.495075(1) | Mixed |
Closest papers by meaning
- One More Time: Revisiting Neural Quantum States from a Reinforcement Learning Perspective (NeurIPS 2026 · similarity 0.88)
- Diagonal Symmetrization of Neural Network Solvers for the Many-Electron Schrödinger Equation (ICML 2025 · similarity 0.88)
- Predicting Ground State Properties: Constant Sample Complexity and Deep Learning Algorithms (NeurIPS 2024 · similarity 0.88)
- A Machine Learning Approach to Duality in Statistical Physics (ICML 2025 · similarity 0.86)
- Score-based Variational Inference via Quantum Maximally Mixed States (NeurIPS 2026 · similarity 0.85)
- PALQO: Physics-informed model for Accelerating Large-scale Quantum Optimization (NeurIPS 2025 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
