Research map / Reinforcement learning / Visual and representation RL
Optimizing Tensor Network Contraction Using Reinforcement Learning
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Use RL-TNCO rather than the strongest reported baseline | Tensor network contraction ordering · Synthetic TNs, n=25, 50, 75, and 100 | mean total flop count: 2×--5× | Helped |
| Use RL-TNCO rather than the strongest reported baseline | Tensor network contraction ordering · Sycamore m=10, 12, 14, and 20 | contraction cost | Mixed |
| Optimize more suffix steps after a greedy prefix | RL-TNCO on Max-cut TN · Max-cut TN | contraction cost: 33.40; 19.05; 9.29 | Helped |
| Add robust feature scaling (ablation) | RL-TNCO · Random TN, 150 nodes | improvement factor over best baseline: \times12.6 -> \boldsymbol{\times44.3} | Helped |
| Add existing-solver information (ablation) | RL-TNCO · Random TN, 150 nodes | improvement factor over best baseline: \times4.5 -> \boldsymbol{\times44.3} | Helped |
| Add an optimistic buffer (ablation) | RL-TNCO · Random TN, 150 nodes | improvement factor over best baseline: \times6.1 -> \boldsymbol{\times44.3} | Helped |
| Add robust feature scaling (ablation) | RL-TNCO · Max-cut TN, last 100 steps | improvement factor over best baseline: \times3.6 -> \times6.2 | Helped |
| Add existing-solver information (ablation) | RL-TNCO · Max-cut TN, last 100 steps | improvement factor over best baseline: \times0.7 -> \times6.2 | Helped |
| Add an optimistic buffer (ablation) | RL-TNCO · Max-cut TN, last 100 steps | improvement factor over best baseline: \times6.2 -> \times6.2 | No clear effect |
| Add path pruning (ablation) | RL-TNCO · Max-cut TN, last 100 steps | improvement factor over best baseline: \times6.2 -> \times6.2 | No clear effect |
| Add robust feature scaling (ablation) | RL-TNCO · Sycamore m=12 | improvement factor over best baseline: \times3.3 -> \boldsymbol{\times7.2} | Helped |
| Add existing-solver information (ablation) | RL-TNCO · Sycamore m=12 | improvement factor over best baseline: \times10^{-3} -> \boldsymbol{\times7.2} | Helped |
Closest papers by meaning
- TensorRL-QAS: Reinforcement learning with tensor networks for improved quantum architecture search (NeurIPS 2025 · similarity 0.91)
- AI-Powered Algorithm-Centric Quantum Processor Topology Design (AAAI 2025 · similarity 0.88)
- One More Time: Revisiting Neural Quantum States from a Reinforcement Learning Perspective (NeurIPS 2026 · similarity 0.87)
- Equivariant Reinforcement Learning for Clifford Quantum Circuit Synthesis (NeurIPS 2026 · similarity 0.87)
- Quantum Positional Encodings for Graph Neural Networks (ICML 2024 · similarity 0.86)
- Curriculum reinforcement learning for quantum architecture search under hardware errors (ICLR 2024 · similarity 0.86)
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.