Research map / Learning theory and algorithms / Matrix and tensor methods
Multi-Stage Multi-Task Learning with Reduced Rank
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 |
|---|---|---|---|
| replace trace norm with capped trace norm (RAMUSA vs MTFL) | multi-task regression · Synthetic omega=15 | MSE: 0.261±0.029 -> 0.212±0.040 | Helped |
| replace trace norm with capped trace norm (RAMUSA vs MTFL) | multi-task regression · Synthetic omega=5 | MSE: 0.210±0.018 -> 0.205±0.021 | No clear effect |
| capped trace norm vs trace norm | multi-task regression · School 10% train | nMSE: 0.570±0.015 -> 0.518±0.012 | Helped |
| capped trace norm vs trace norm | multi-task regression · SARCOS 300 train | nMSE: 0.136±0.003 -> 0.137±0.004 | No clear effect |
| capped trace norm vs trace norm | multi-task regression · Microarray 20% train | nMSE: 0.746±0.038 -> 0.739±0.046 | No clear effect |
| capped trace norm vs trace norm | multi-task regression · Traffic 20% train | nMSE: 0.328±0.006 -> 0.316±0.006 | Helped |
| capped trace norm vs trace norm | multi-task classification · Handwritten Letter 20% train | test error (%): 11.86±6.49 -> 8.48±6.08 | Helped |
| multi-task sharing (MTFL) vs single-task ridge regression | multi-task regression · School 10% train | nMSE: 1.854±0.026 -> 0.570±0.015 | Helped |
| vary threshold tau from 0 upward (ablation) | RAMUSA · Synthetic data | MSE | Mixed |
| increase number of stages (ablation) | RAMUSA · Synthetic data | estimation error | Helped |
Closest papers by meaning
- Near-Optimal and Efficient First-Order Algorithm for Multi-Task Learning with Shared Linear Representation (ICML 2026 · similarity 0.87)
- Task Groupings Regularization: Data-Free Meta-Learning with Heterogeneous Pre-trained Models (ICML 2024 · similarity 0.85)
- Robust Multi-Task Learning with Excess Risks (ICML 2024 · similarity 0.85)
- MTLoRA: Low-Rank Adaptation Approach for Efficient Multi-Task Learning (CVPR 2024 · similarity 0.85)
- Towards Principled Task Grouping for Multi-Task Learning (NeurIPS 2026 · similarity 0.84)
- NTKMTL: Mitigating Task Imbalance in Multi-Task Learning from Neural Tangent Kernel Perspective (NeurIPS 2025 · similarity 0.84)
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.