Research map / Graph learning / Knowledge and temporal graphs
Attentive Relational Networks for Mapping Images to Scene Graphs
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 semantic transformation and graph self-attention to the baseline (ablation) | Attentive Relational Network · Visual Genome test, SGCls | Recall@50: 34.6 -> 38.2 | Helped |
| add graph self-attention to the baseline (ablation) | Attentive Relational Network · Visual Genome test, SGCls | Recall@50: 34.6 -> 37.2 | Helped |
| add semantic transformation to the baseline (ablation) | Attentive Relational Network · Visual Genome test, SGCls | Recall@50: 34.6 -> 37.3 | Helped |
| add semantic transformation to graph self-attention (ablation) | Attentive Relational Network · Visual Genome test, SGCls | Recall@50: 37.2 -> 38.2 | Helped |
| add graph self-attention to semantic transformation (ablation) | Attentive Relational Network · Visual Genome test, SGCls | Recall@50: 37.3 -> 38.2 | Helped |
| remove semantic loss and use summation instead of concatenation (ablation) | Attentive Relational Network · Visual Genome test, SGCls | Recall@50: 38.2 -> 36.6 | Hurt |
| remove semantic loss and use element-wise multiplication (ablation) | Attentive Relational Network · Visual Genome test, SGCls | Recall@50: 38.2 -> 34.0 | Hurt |
| remove semantic loss while retaining concatenation (ablation) | Attentive Relational Network · Visual Genome test, SGCls | Recall@50: 38.2 -> 36.2 | Hurt |
| replace concatenation with summation (ablation) | Attentive Relational Network · Visual Genome test, SGCls | Recall@50: 38.2 -> 36.9 | Hurt |
| replace concatenation with element-wise multiplication (ablation) | Attentive Relational Network · Visual Genome test, SGCls | Recall@50: 38.2 -> 36.6 | Hurt |
| remove word embeddings from semantic transformation (ablation) | Attentive Relational Network · Visual Genome test, SGCls | Recall@50: 38.2 -> 37.3 | Hurt |
| replace multi-head attention with single-head attention (ablation) | Attentive Relational Network · Visual Genome test, SGCls | Recall@50: 38.2 -> 37.9 | Hurt |
Closest papers by meaning
- LinkNet: Relational Embedding for Scene Graph (NeurIPS 2018 · similarity 0.92)
- Scene Graph Generation by Iterative Message Passing (CVPR 2017 · similarity 0.91)
- DSGG: Dense Relation Transformer for an End-to-end Scene Graph Generation (CVPR 2024 · similarity 0.90)
- Joint Modeling of Visual Objects and Relations for Scene Graph Generation (NeurIPS 2021 · similarity 0.89)
- EGTR: Extracting Graph from Transformer for Scene Graph Generation (CVPR 2024 · similarity 0.89)
- Scene Graph-Grounded Image Generation (AAAI 2025 · similarity 0.87)
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.