Research map / Vision-language and multimodal / Visual question answering
DualVD: An Adaptive Dual Encoding Model for Deep Visual Understanding in Visual Dialogue
Xiaoze Jiang, Jing Yu, Zengchang Qin, Yingying Zhuang, Xingxing Zhang, Yue Hu, Qi Wu
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 |
|---|---|---|---|
| combine appearance-level visual features with semantic captions using adaptive fusion (ablation) | DualVD · VisDial v1.0 validation | MRR: 61.07 -> 64.64 | Improved |
| add relational semantics to object and neighborhood features (ablation) | DualVD visual module · VisDial v1.0 validation | MRR: 63.97 -> 64.11 | Improved |
| combine object appearance with relation-aware object features (ablation) | DualVD visual module · VisDial v1.0 validation | MRR: 63.84 -> 64.11 | Improved |
| use relationally enhanced object representations instead of object features alone (ablation) | DualVD visual module · VisDial v1.0 validation | MRR: 63.84 -> 64.11 | Improved |
| fuse global and local caption information instead of using either alone (ablation) | DualVD semantic module · VisDial v1.0 validation | MRR: 60.02 -> 61.07; 60.95 -> 61.07 | Improved |
| add ELMo word embeddings (ablation) | DualVD · VisDial v1.0 validation | MRR: 63.67 -> 64.64 | Improved |
| strengthen relational features while weakening object appearance (ablation) | DualVD visual representation · VisDial v1.0 validation | MRR: 63.63 vs 64.11 | Hurt |
| use global and local captions together rather than global captions alone (ablation) | DualVD semantic module · VisDial v1.0 validation | MRR: 60.02 -> 61.07 | Improved |
| use global and local captions together rather than local captions alone (ablation) | DualVD semantic module · VisDial v1.0 validation | MRR: 60.95 -> 61.07 | Improved |
| use DualVD rather than RvA | VisDial v1.0 · test-standard | MRR: 63.03 -> 63.23 | Improved |
| use DualVD rather than DL-61 | VisDial v1.0 · test-standard | NDCG: 57.32 -> 56.32 | Hurt |
Closest papers by meaning
- Multi-Level Attention Networks for Visual Question Answering (CVPR 2017 · similarity 0.90)
- AdaptVision: Efficient Vision-Language Models via Adaptive Visual Acquisition (CVPR 2026 · similarity 0.89)
- V^2Dial: Unification of Video and Visual Dialog via Multimodal Experts (CVPR 2025 · similarity 0.89)
- Language-aware Visual Semantic Distillation for Video Question Answering (CVPR 2024 · similarity 0.89)
- Object Attribute Matters in Visual Question Answering (AAAI 2024 · similarity 0.89)
- ConVQG: Contrastive Visual Question Generation with Multimodal Guidance (AAAI 2024 · similarity 0.89)
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.
