Research map / Label-efficient and robust learning / Recognition and long-tail
TS-MOF: Two-Stage Multi-Objective Fine-tuning for Long-Tailed Recognition
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's abstract only. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Proposes TS-MOF, a two-stage multi-objective fine-tuning framework for long-tailed recognition that decouples the problem to make multi-objective optimization over multiple LTR strategies more stable.
Problem
Long-tailed recognition under extreme class imbalance, where methods struggle to balance head and tail class performance; directly applying multi-objective optimization to combine multiple LTR strategies is complex and unstable.
Approach
Strategically decomposes the learning into two stages with multi-objective fine-tuning, rather than applying multi-objective optimization over multiple LTR strategies in a single end-to-end step.
Closest papers by meaning
- U-MOF: Uncertainty-Guided Parameter-Efficient Multi-Objective Fine-Tuning for Long-Tailed Recognition (NeurIPS 2026 · similarity 0.96)
- Long-Tailed Learning as Multi-Objective Optimization (AAAI 2024 · similarity 0.88)
- Feature Fusion from Head to Tail for Long-Tailed Visual Recognition (AAAI 2024 · similarity 0.87)
- Long-tailed Recognition with Model Rebalancing (NeurIPS 2025 · similarity 0.87)
- Decoupled Optimisation for Long-Tailed Visual Recognition (AAAI 2024 · similarity 0.87)
- Long-Tailed Recognition via Information-Preservable Two-Stage Learning (NeurIPS 2025 · similarity 0.86)
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.