Research map / Label-efficient and robust learning / Recognition and long-tail
Multiplicative Multitask Feature Learning
Xin Wang, Jinbo Bi, Shipeng Yu, Jiangwen Sun, Minghu Song
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 MMTFL(2,1) (l2 on c, l1 on beta) instead of MMTFL(2,2) | multiplicative MTFL · Synthetic D1 (33% training data) | R2: 0.79 -> 0.86 | Improved |
| use MMTFL(2,1) instead of MMTFL(1,1) | multiplicative MTFL · Synthetic D1 (25% training data) | R2: 0.54 -> 0.73 | Improved |
| use MMTFL(1,2) instead of MMTFL(2,2) | multiplicative MTFL · Synthetic D2 (33% training data) | R2: 0.69 -> 0.83 | Improved |
| use MMTFL(1,2) instead of MMTFL(2,1) | multiplicative MTFL · Synthetic D1 (25% training data) | R2: 0.73 -> 0.42 | Hurt |
| use multiplicative MMTFL(2,1) instead of additive dirty model DMTL | multitask feature learning · Synthetic D1 (33% training data) | R2: 0.73 -> 0.86 | Improved |
| use multiplicative MMTFL(1,2) instead of additive robust MTFL | multitask feature learning · Synthetic D2 (50% training data) | R2: 0.83 -> 0.97 | Improved |
| use multitask learning DMTL instead of single task learning | single task learning · SARCOS (25% training data) | R2: 0.78 -> 0.90 | Improved |
| use MMTFL(1,2) instead of MMTFL(2,2) | multiplicative MTFL · SARCOS (33% training data) | R2: 0.90 -> 0.89 | Hurt |
| use MMTFL(1,2) instead of rMTFL | multitask feature learning · USPS (33% training data) | F1 score: 0.90 -> 0.91 | Improved |
| use MMTFL(2,1) instead of DMTL | multitask feature learning · SARCOS (50% training data) | R2: 0.87 -> 0.91 | Improved |
| increase training data fraction from 25% to 50% for MMTFL(1,2) | MMTFL(1,2) · Synthetic D2 | R2: 0.49 -> 0.97 | Improved |
Closest papers by meaning
- The Benefit of Multitask Representation Learning (JMLR 2016 · similarity 0.88)
- Consistent Multitask Learning with Nonlinear Output Relations (NeurIPS 2017 · similarity 0.88)
- Learning Sparse Task Relations in Multi-Task Learning (AAAI 2017 · similarity 0.87)
- Provable Multi-Task Representation Learning by Two-Layer ReLU Neural Networks (ICML 2024 · similarity 0.86)
- Identifying and Mitigating Spurious Correlation in Multi-Task Learning (CVPR 2025 · similarity 0.86)
- Joint-Task Regularization for Partially Labeled Multi-Task Learning (CVPR 2024 · 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.
