Research map / Learning theory and algorithms / Matrix and tensor methods
Nonnegative Tensor Completion via Integer Optimization
Caleb Xavier Bugg, Chen Chen, Anil Aswani
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 |
|---|---|---|---|
| use BCG with nonnegative-tensor-norm vs ALS | nonnegative tensor completion · order-3 10x10x10, n=500 | NMSE: 0.058 -> 0.004 | Improved |
| use BCG vs ALS as tensors grow to 100x100x100 with n=500 | nonnegative tensor completion · order-3 100x100x100 | NMSE: 0.666 -> 0.641 | No clear effect |
| BCG vs ALS at larger sizes (80x80x80) | nonnegative tensor completion · order-3 80x80x80, n=500 | NMSE: 0.539 -> 0.554 | Hurt |
| use BCG vs ALS as order increases | nonnegative tensor completion · 10^8 entries, n=10,000 | NMSE: 0.986 -> 0.156 | Improved |
| use BCG vs SiLRTC and TNCP baselines | nonnegative tensor completion · 10^6 entries, n=10,000 | NMSE: 0.883 -> 0.008 | Improved |
| increase sample percent from 0.01% to 1% | BCG · 10^6 tensor | NMSE: 0.989 -> 0.007 | Improved |
| increase sample percent from 1% to 10% | BCG · 10^6 tensor | NMSE: 0.007 -> 0.005 | No clear effect |
| increase sample percent from 1% to 10% | BCG · 10^7 tensor | NMSE: 0.011 -> 0.013 | No clear effect |
| use BCG vs ALS with 1% samples | nonnegative tensor completion · 10^7 tensor | NMSE: 0.220 -> 0.011 | Improved |
| use BCG vs ALS in compute time | nonnegative tensor completion · 10^8 entries, n=10,000 | time (s): 15.1 -> 872 | Hurt |
| use alternating maximization heuristic before integer program in separation oracle (ablation) | BCG weak oracle · runtime | time | Improved |
Closest papers by meaning
- A Dual Framework for Low-rank Tensor Completion (NeurIPS 2018 · similarity 0.91)
- Tensor Completion Made Practical (NeurIPS 2020 · similarity 0.90)
- Union of Low-Rank Tensor Spaces: Clustering and Completion (JMLR 2020 · similarity 0.90)
- Fast Tensor Completion via Approximate Richardson Iteration (ICML 2025 · similarity 0.89)
- Generalization Analysis of Deep Non-linear Matrix Completion (ICML 2024 · similarity 0.87)
- Symmetric Matrix Completion with ReLU Sampling (ICML 2024 · similarity 0.87)
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.
