Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
MTNL: A Unified Modeling Perspective for Enhancing Tensor Network Learning
Junhua Zeng, Yuning Qiu, Binghua Li, Chao Li, Qibin Zhao, Guoxu Zhou
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 |
|---|---|---|---|
| fuse multiple TNs in a deep-network style (MTNL) instead of single-structure TN completion | MTNL · CAVE MSI completion, average RSE 85% missing | RSE: 0.0465 -> 0.0414 vs BayesTNSS | Improved |
| fuse multiple TNs in a deep-network style (MTNL) instead of single-structure TN completion | MTNL · CAVE MSI completion, average RSE 95% missing | RSE: 0.1211 -> 0.1143 vs BTT | Improved |
| rank-cut threshold 1e-1 and retrain the searched MTNL structure (ablation) | MTNL · DROP, LLaMA2-7B | F1: 59.6 -> 60.1 | Improved |
| learned MTNL structure vs QuanTA at similar parameter count | MTNL (rank-cut) · DROP, LLaMA2-7B | F1: 59.5 -> 60.1 | Improved |
| add MTNL-dropout post-processing for uncertainty | MTNL (4,4) regression · Tensor regression under distribution shift 0 to 9 | prediction variance: 0.0001 -> 0.0017 | Improved |
| MTNL-dropout instead of SVGD for uncertainty | MTNL-dropout · Tensor regression 64x64, NLL | compute cost: up to 200x lower cost | Improved |
| add core prior to TN-SS (ablation) | MTNL · Synthetic TN-SS D=6 | learned parameters: 248 -> 178 | Improved |
| add threshold prior to TN-SS (ablation) | MTNL · Synthetic TN-SS D=8 | learned parameters: 454 -> 328 | Improved |
| add core prior (ablation) | MTNL · Synthetic TN-SS D=6 | iterations required (x1000): 3215 -> 655 | Improved |
| add threshold prior (ablation) | MTNL · Synthetic TN-SS D=8 | iterations required (x1000): 8125 -> 460 | Improved |
Closest papers by meaning
- tnGPS: Discovering Unknown Tensor Network Structure Search Algorithms via Large Language Models (LLMs) (ICML 2024 · similarity 0.88)
- SVDinsTN: A Tensor Network Paradigm for Efficient Structure Search from Regularized Modeling Perspective (CVPR 2024 · similarity 0.87)
- Convolutions and More as Einsum: A Tensor Network Perspective with Advances for Second-Order Methods (NeurIPS 2024 · similarity 0.87)
- Factor Augmented Tensor-on-Tensor Neural Networks (AAAI 2025 · similarity 0.86)
- Deep Tree Tensor Networks (NeurIPS 2025 · similarity 0.86)
- Geometry-aware training of factorized layers in tensor Tucker format (NeurIPS 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.
