Research map / Vision-language and multimodal / Human and embodied modeling
Stacked Hybrid-Attention and Group Collaborative Learning for Unbiased Scene Graph Generation
Check 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 |
|---|---|---|---|
| Add GCL to SHA | SHA · VG150 PredCls | mR@100: 20.5 -> 44.0 | Helped |
| Add GCL to SHA | SHA · GQA200 PredCls | mR@100: 21.1 -> 42.7 | Helped |
| Add GCL to Motifs and VCTree | Motifs and VCTree · VG150, GQA200 PredCls/SGCls/SGDet | mR@100 | Helped |
| Add SHA hybrid attention to the encoder (ablation) | SHA + GCL · VG150 PredCls | mR@100: 41.7 -> 44.0 | Helped |
| Add parallel classifier optimization (PCO) (ablation) | GCL · VG150 PredCls | mR@100: 37.4 -> 41.7 | Helped |
| Add collaborative knowledge distillation (CKD) (ablation) | GCL · VG150 PredCls | mR@100: 41.7 -> 44.0 | Helped |
| Add GCL instead of using a one-layer relation classifier (ablation) | SHA relation decoder · VG150 PredCls | mR@100: 20.5 -> 44.0 | Helped |
| Add cross-attention to SHA layers (ablation) | SHA + GCL · VG150 PredCls | mR@100: 42.5 -> 44.0 | Helped |
| Add self-attention to SHA layers (ablation) | SHA + GCL · VG150 PredCls | mR@100: 41.5 -> 44.0 | Helped |
| Use μ=4 rather than μ=3 or μ=5 for predicate grouping (ablation) | GCL · VG150 PredCls | mR@100: 42.4, 43.5, 41.7 -> 43.2, 44.0, 42.0 | Mixed |
| Use Top-down rather than Adjacent knowledge matching (ablation) | GCL · VG150 PredCls | mR@100: 43.5 -> 44.0 | Helped |
Closest papers by meaning
- Adaptive Self-training Framework for Fine-grained Scene Graph Generation (ICLR 2024 · similarity 0.89)
- Navigating the Unseen: Zero-shot Scene Graph Generation via Capsule-Based Equivariant Features (CVPR 2025 · similarity 0.89)
- Universal Scene Graph Generation (CVPR 2025 · similarity 0.88)
- RA-SGG: Retrieval-Augmented Scene Graph Generation Framework via Multi-Prototype Learning (AAAI 2025 · similarity 0.88)
- Panoptic Scene Graph Generation with Semantics-Prototype Learning (AAAI 2024 · similarity 0.88)
- LLM4SGG: Large Language Models for Weakly Supervised Scene Graph Generation (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.