Research map / Label-efficient and robust learning / Noisy labels
NICEST: Noisy Label Correction and Training for Robust Scene Graph Generation
Lin Li, Jun Xiao, Hanrong Shi, Hanwang Zhang, Yi Yang, Wei Liu, Long Chen
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 |
|---|---|---|---|
| add Neg-NSD (mine missing negative samples) (ablation) | Motifs · VG PredCls | Mean: 41.8 -> 43.8 | Improved |
| add Pos-NSD alone (ablation) | Motifs · VG PredCls | Mean: 41.8 -> 39.8 | Hurt |
| add Neg-NSD and Pos-NSD without NSC (ablation) | Motifs · VG PredCls | Mean: 41.8 -> 40.5 | Hurt |
| add NSC on top of Pos-NSD (ablation) | Motifs · VG PredCls | Mean: 39.8 -> 42.4 | Improved |
| add Neg-NSD to Pos-NSD + NSC (ablation) | Motifs · VG PredCls | Mean: 42.4 -> 44.3 | Improved |
| use full NICE vs baseline | Motifs · VG PredCls | mR@100: 17.8 -> 32.1 | Improved |
| add NIST multi-teacher distillation on NICE | Motifs+NICE · VG PredCls | Mean: 44.3 -> 45.3 | Improved |
| add NIST on NICE | Motifs+NICE · VG PredCls | R@100: 58.6 -> 61.0 | Improved |
| add NIST on NICE | Motifs+NICE · VG PredCls | mR@100: 32.1 -> 31.6 | Hurt |
| add frequency prior (ablation) | Motifs · VG-OOD PredCls | Mean: 34.9 -> 36.5 without prior | Hurt |
| add frequency prior (ablation) | Motifs · VG PredCls | Mean: 40.9 -> 41.8 | Improved |
| NICE/NICEST vs label correction and label smoothing | Motifs · VG PredCls | Mean: 3.9% ~ 4.7% absolute gains | Improved |
Closest papers by meaning
- The Devil Is in the Labels: Noisy Label Correction for Robust Scene Graph Generation (CVPR 2022 · similarity 0.96)
- Adaptive Self-training Framework for Fine-grained Scene Graph Generation (ICLR 2024 · similarity 0.88)
- Semi-Supervised Clustering Framework for Fine-grained Scene Graph Generation (AAAI 2025 · similarity 0.88)
- TD²-Net: Toward Denoising and Debiasing for Video Scene Graph Generation (AAAI 2024 · similarity 0.88)
- RA-SGG: Retrieval-Augmented Scene Graph Generation Framework via Multi-Prototype Learning (AAAI 2025 · similarity 0.87)
- Scene Graph Generation Strategy with Co-occurrence Knowledge and Learnable Term Frequency (ICML 2024 · 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.
