Research map / Vision-language and multimodal / Visual question answering
Relation-Aware Graph Attention Network for Visual Question Answering
Linjie Li, Zhe Gan, Yu Cheng, Jingjing 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 attention (GAT instead of GCN) to semantic relation encoder (ablation) | RaAM Sem (BUTD) · VQA 2.0 val | accuracy: +0.70 | Improved |
| add attention (GAT instead of GCN) to spatial relation encoder (ablation) | RaAM Spa (BUTD) · VQA 2.0 val | accuracy: +0.81 | Improved |
| add question-adaptive concatenation (no attention) (ablation) | RaAM Sem/Spa (BUTD) · VQA 2.0 val | accuracy: +0.1 | Improved |
| add question-adaptive concatenation (with attention) to semantic (ablation) | RaAM Sem (BUTD) · VQA 2.0 val | accuracy: 63.90 -> 64.11 | Improved |
| add question-adaptive concatenation (with attention) to implicit (ablation) | RaAM Imp (BUTD) · VQA 2.0 val | accuracy: 63.36 -> 64.10 | Improved |
| combine implicit, semantic and spatial relations | BUTD · VQA 2.0 val | accuracy: 63.38 -> 65.30 | Improved |
| combine implicit, semantic and spatial relations | MUTAN · VQA 2.0 val | accuracy: 61.36 -> 64.37 | Improved |
| combine implicit, semantic and spatial relations | BAN · VQA 2.0 val | accuracy: 65.51 -> 67.18 | Improved |
| add spatial relation encoder | BAN (four glimpses) · VQA-CP v2 test | accuracy: 39.24 -> 40.30 | Improved |
| add semantic relation encoder | BAN (four glimpses) · VQA-CP v2 test | accuracy: 39.24 -> 39.54 | Improved |
| add implicit relation encoder | BAN (four glimpses) · VQA-CP v2 test | accuracy: 39.24 -> 39.58 | Improved |
| add semantic relation encoder | MUTAN · VQA 2.0 val | accuracy: 61.36 -> 62.60 | Improved |
Closest papers by meaning
- MRA-Net: Improving VQA Via Multi-Modal Relation Attention Network (TPAMI 2020 · similarity 0.95)
- Graph-Structured Representations for Visual Question Answering (CVPR 2017 · similarity 0.92)
- Dynamic Key-Value Memory Enhanced Multi-Step Graph Reasoning for Knowledge-Based Visual Question Answering (AAAI 2022 · similarity 0.92)
- Structured Attentions for Visual Question Answering (ICCV 2017 · similarity 0.92)
- Multi-Level Attention Networks for Visual Question Answering (CVPR 2017 · similarity 0.92)
- Object Attribute Matters in Visual Question Answering (AAAI 2024 · similarity 0.90)
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.
