Research map / Learning theory and algorithms / Matrix and tensor methods
Scalable Interpretable Multi-Response Regression via SEED
Zemin Zheng, M. Taha Bahadori, Yan Liu, Jinchi Lv
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 SEED instead of LN-ADMM and RCGL | LN-ADMM / RCGL · Simulation 1, p=400, r=3 | normalized prediction error (x1e-2): 0.8140 vs 13.5766 and 29.7221 | Improved |
| use SEED instead of nuclear-norm penalization (LN-ADMM) | LN-ADMM · Simulation 1, p=2000, r=3 | rank recovery error: 2.8700 -> 0.0000 | Improved |
| use SEED instead of RCGL iterative SVD | RCGL · Simulation 1, p=2000, r=3 | rank recovery error: 2.5100 -> 0.0000 | Improved |
| use SEED instead of LN-ADMM for support recovery at p=2000, r=3 | LN-ADMM · Simulation 1, p=2000, r=3 | support recovery AUC: 0.9962 vs 0.9738 | Hurt |
| use SEED for estimation at p=100, r=30 | LN-ADMM · Simulation 1, p=100, r=30 | normalized estimation error (x1e-2): 12.4132 -> 2.2812 | Improved |
| use SEED instead of baselines for CPU runtime | LN-ADMM and RCGL · Simulation 2, CPU (n=q=1000, r=1) | runtime: speed up of 10-100 times | Improved |
| run SEED with the two-step eigen-decomposition on GPU | SEED · Simulation 2, GPU (q=10000, n=5000, r=1) | runtime: estimate a matrix with 10^8 elements in less than a minute | Improved |
| use SEED instead of baselines for graph recovery | LN-ADMM and RCGL · Haiti Twitter influence network | AUC vs retweet network: SEED best at rank 4 | Improved |
| increase solution rank in influence network inference (ablation) | all algorithms · Haiti Twitter influence network | graph recovery AUC: improves initially then saturates | Mixed |
| use SEED for run time on the application data | LN-ADMM and RCGL · Haiti Twitter data | run time (seconds): 127.34 and 256.87 -> 2.83 | Improved |
Closest papers by meaning
- Fast Stagewise Sparse Factor Regression (JMLR 2022 · similarity 0.87)
- Communication-efficient Sparse Regression (JMLR 2017 · similarity 0.85)
- On $b$-bit Min-wise Hashing for Large-scale Regression and Classification with Sparse Data (JMLR 2018 · similarity 0.85)
- A Decomposition Algorithm for the Sparse Generalized Eigenvalue Problem (CVPR 2019 · similarity 0.85)
- Low-rank Interaction with Sparse Additive Effects Model for Large Data Frames (NeurIPS 2018 · similarity 0.85)
- Spectral Deconfounding via Perturbed Sparse Linear Models (JMLR 2020 · 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.
