AI Paper Atlas

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

ICML 2026 · Probabilistic and scientific modeling › Uncertainty and Gaussian processes · approach: Priors and functional variational uncertainty

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 changedWhereMeasuredEffect
fuse multiple TNs in a deep-network style (MTNL) instead of single-structure TN completionMTNL · CAVE MSI completion, average RSE 85% missingRSE: 0.0465 -> 0.0414 vs BayesTNSSImproved
fuse multiple TNs in a deep-network style (MTNL) instead of single-structure TN completionMTNL · CAVE MSI completion, average RSE 95% missingRSE: 0.1211 -> 0.1143 vs BTTImproved
rank-cut threshold 1e-1 and retrain the searched MTNL structure (ablation)MTNL · DROP, LLaMA2-7BF1: 59.6 -> 60.1Improved
learned MTNL structure vs QuanTA at similar parameter countMTNL (rank-cut) · DROP, LLaMA2-7BF1: 59.5 -> 60.1Improved
add MTNL-dropout post-processing for uncertaintyMTNL (4,4) regression · Tensor regression under distribution shift 0 to 9prediction variance: 0.0001 -> 0.0017Improved
MTNL-dropout instead of SVGD for uncertaintyMTNL-dropout · Tensor regression 64x64, NLLcompute cost: up to 200x lower costImproved
add core prior to TN-SS (ablation)MTNL · Synthetic TN-SS D=6learned parameters: 248 -> 178Improved
add threshold prior to TN-SS (ablation)MTNL · Synthetic TN-SS D=8learned parameters: 454 -> 328Improved
add core prior (ablation)MTNL · Synthetic TN-SS D=6iterations required (x1000): 3215 -> 655Improved
add threshold prior (ablation)MTNL · Synthetic TN-SS D=8iterations required (x1000): 8125 -> 460Improved

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy