Research map / Optimization / Optimizers
Multi-Task Learning with User Preferences: Gradient Descent with Controlled Ascent in Pareto Optimization
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 |
|---|---|---|---|
| Controlled ascent combined with multiple gradient descent (EPO Search) vs. PMTL with reference vectors | MTL on MultiMNIST, MultiFashion, Multi-Fashion+MNIST · LeNet MTL classification | per-task test accuracy | Helped |
| Linear scalarization with preference weights vs. MOO-based methods (EPO, PMTL) | MTL classification and regression · MultiMNIST etc. and River Flow | accuracy and RLP | Hurt |
| EPO Search vs. PMTL on far-initialized synthetic problem | Synthetic two-objective problem (36) · n=20 synthetic | reaching Pareto front | Helped |
| EPO Search vs. PMTL on 8-task River Flow regression | FNN regression on river flow · River Flow (m=8) | relative loss profile (lower better) | Helped |
| EPO Search vs. single-task baselines on River Flow | FNN regression on river flow · River Flow (m=8) | relative loss profile (lower better) | Helped |
Closest papers by meaning
- Delve into the Applicability of Advanced Optimizers for Multi-Task Learning (NeurIPS 2026 · similarity 0.91)
- Towards Consistent Multi-Task Learning: Unlocking the Potential of Task-Specific Parameters (CVPR 2025 · similarity 0.90)
- Fair Resource Allocation in Multi-Task Learning (ICML 2024 · similarity 0.89)
- Bayesian Uncertainty for Gradient Aggregation in Multi-Task Learning (ICML 2024 · similarity 0.88)
- Towards Principled Task Grouping for Multi-Task Learning (NeurIPS 2026 · similarity 0.88)
- MTLoRA: Low-Rank Adaptation Approach for Efficient Multi-Task Learning (CVPR 2024 · similarity 0.88)
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.