Research map / Label-efficient and robust learning / Semi-supervised learning
Rethinking Low-Confidence Pseudo Labels: Influence-Aware Semi-Supervised Fine-Tuning for Hyperspectral Change Detection
Keyun Zhao, Guangchang Li, Yunpeng Bai, Jiang Shao, Ying Li
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 LoRA fine-tuning instead of full fine-tuning (ablation) | SpatialSigma foundation model · River | Kappa: 72.97 -> 76.64 | Improved |
| add IA-SFT (ablation) | LoRA FT baseline · River | Kappa: 76.64 -> 81.31 | Improved |
| add AFCD decoder (ablation) | LoRA FT baseline · River | Kappa: 76.64 -> 80.77 | Improved |
| add both IA-SFT and AFCD (ablation) | LoRA FT baseline · River | Kappa: 76.64 -> 83.11 | Improved |
| add AFCD decoder (without IA-SFT) (ablation) | LoRA FT baseline · River | Recall: 73.16 -> 74.35 recall but precision 84.72 -> 92.05 | Hurt |
| lower selection ratio of pseudo-labels by mean-variance influence (ablation) | IA-SFT · River, China, USA | pseudo-label error rate | Improved |
| select pseudo-labels with IA-SFT instead of FixMatch/FlexMatch/CGMatch (ablation) | IAF-CDNet · River | pseudo-label accuracy: 86.52 -> 94.73 (vs FixMatch) | Improved |
| use eFUMI for pseudo-label selection (ablation) | IAF-CDNet · China, USA | pseudo-label accuracy: USA PLA 88.94 vs IA-SFT 95.12 | Mixed |
| plug IA-SFT into AIWSEN (ablation) | AIWSEN · USA | Kappa: 77.96 -> 80.39 | Improved |
| plug IA-SFT into SpikeHCD (ablation) | SpikeHCD · China | Kappa: 89.83 -> 92.11 | Improved |
Closest papers by meaning
- Single-Temporal Supervised Learning for Universal Remote Sensing Change Detection (IJCV 2024 · similarity 0.86)
- Feature Spectrum Learning for Remote Sensing Change Detection (CVPR 2025 · similarity 0.85)
- Change Is Everywhere: Single-Temporal Supervised Object Change Detection in Remote Sensing Imagery (ICCV 2021 · similarity 0.85)
- RSRCC: A Remote Sensing Regional Change Comprehension Benchmark Constructed via Retrieval-Augmented Best-of-𝑁 Ranking (NeurIPS 2026 · similarity 0.83)
- The Change You Want To Detect: Semantic Change Detection In Earth Observation With Hybrid Data Generationf (CVPR 2025 · similarity 0.83)
- Meta-Learning Hyperparameters for Parameter Efficient Fine-Tuning (CVPR 2025 · similarity 0.83)
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.
