Research map / Probabilistic and scientific modeling / Causal discovery
A Causal Adjustment Module for Debiasing Scene Graph Generation
Li Liu, Shuzhou Sun, Shuaifeng Zhi, Fan Shi, Zhen Liu, Janne Heikkilä, Yongxiang Liu
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 CAModule | MotifsNet · VG150 PredCls | mR@100: 16.8 -> 39.3 | Improved |
| add CAModule | MotifsNet · VG150 SGCls | mR@100: 9.5 -> 24.7 | Improved |
| add CAModule | MotifsNet · VG150 SGDet | mR@100: 8.5 -> 18.2 | Improved |
| add CAModule | VCTree · VG150 PredCls | mR@100: 16.6 -> 40.5 | Improved |
| add CAModule | VCTree · VG150 SGCls | mR@100: 8.0 -> 25.9 | Improved |
| add CAModule | Transformer · VG150 PredCls | mR@100: 17.5 -> 39.4 | Improved |
| add CAModule | Transformer · VG150 SGDet | mR@100: 8.8 -> 19.3 | Improved |
| add CAModule | MotifsNet · GQA PredCls | mR@100: 16.3 -> 38.7 | Improved |
| add CAModule | Transformer · GQA PredCls | mR@100: 17.7 -> 40.2 | Improved |
| use CAModule instead of GCL as debiasing method | MotifsNet · VG150 SGDet | mR@100: 19.3 -> 18.2 | Hurt |
| use CAModule instead of FGPL as debiasing method | VCTree · VG150 SGCls | mR@100: 27.6 -> 25.9 | Hurt |
| use CAModule instead of GCL as debiasing method | MotifsNet · VG150 PredCls | mR@100: 38.2 -> 39.3 | Improved |
Closest papers by meaning
- Unbiased Scene Graph Generation via Two-Stage Causal Modeling (TPAMI 2023 · similarity 0.94)
- TD²-Net: Toward Denoising and Debiasing for Video Scene Graph Generation (AAAI 2024 · similarity 0.87)
- Towards Unbiased and Robust Spatio-Temporal Scene Graph Generation and Anticipation (CVPR 2025 · similarity 0.87)
- Multimodal Event Causality Reasoning with Scene Graph Enhanced Interaction Network (AAAI 2024 · similarity 0.86)
- A Closer Look at Dynamic Scene Graph Generation in the Era of Multimodal Large Language Models (NeurIPS 2026 · similarity 0.86)
- CausalVerse: Benchmarking Causal Representation Learning with Configurable High-Fidelity Simulations (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.
