Research map / Graph learning / Knowledge and temporal graphs
Iterative Context-Aware Graph Inference for Visual Dialog
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 |
|---|---|---|---|
| Use 3 inference steps instead of 1 (ablation) | CAG · VisDial v0.9 val | R@1: 53.25 -> 54.64 | Helped |
| Use 4 inference steps instead of 3 (ablation) | CAG · VisDial v0.9 val | R@1: 54.64 -> 54.11 | Hurt |
| Add dynamic directed-graph inference (ablation) | CAG · VisDial v0.9 val | MRR: 65.73 -> 67.56 | Helped |
| Add history context u (ablation) | CAG · VisDial v0.9 val | R@1: 51.83 -> 54.64 | Helped |
| Use word-level question attention instead of sentence-level question feature (ablation) | CAG · VisDial v0.9 val | R@1: 53.74 -> 54.64 | Helped |
| Add graph attention instead of average-pooling node representations (ablation) | CAG · VisDial v0.9 val | R@1: 53.99 -> 54.64 | Helped |
| Use asymmetric question guidance on one node side in adjacency correlation (ablation) | CAG · VisDial v0.9 val | R@1: 54.16 -> 54.64 | Helped |
| Set neighbor count K to 8 rather than fewer or more neighbors (ablation) | CAG · VisDial v0.9 val | R@1 | Mixed |
| Use object-region features rather than VGG features | CAG · VisDial v0.9 val | R@1: 51.45 -> 54.64 | Helped |
| Use CAG rather than GNN | Visual dialog · VisDial v0.9 val | R@1: 48.95 -> 54.64 | Helped |
| Use CAG rather than FGA | Visual dialog · VisDial v0.9 val | R@1: 51.43 -> 54.64 | Helped |
| Use CAG rather than GNN | Visual dialog · VisDial v1.0 test-std | NDCG: 52.82 -> 56.64 | Helped |
Closest papers by meaning
- Structure-Aware Multimodal Sequential Learning for Visual Dialog (AAAI 2024 · similarity 0.89)
- Visual Graph Arena: Evaluating Visual Conceptualization of Vision and Multimodal Large Language Models (ICML 2025 · similarity 0.88)
- Relation-aware Hierarchical Prompt for Open-vocabulary Scene Graph Generation (AAAI 2025 · similarity 0.86)
- Scene Graph Thinking: Reinforcing Structured Visual Reasoning for Multimodal Large Language Models (ICML 2026 · similarity 0.86)
- Talk like a Graph: Encoding Graphs for Large Language Models (ICLR 2024 · similarity 0.86)
- GRAPHGPT-O: Synergistic Multimodal Comprehension and Generation on Graphs (CVPR 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.