Research map / Vision-language and multimodal / Visual question answering
Dynamic Key-Value Memory Enhanced Multi-Step Graph Reasoning for Knowledge-Based Visual Question Answering
Read the paperCheck 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 dense captioning information (ablation) | DMMGR · KRVQR | top-1 accuracy: 31.4 -> 31.8 | No clear effect |
| add dense captioning information (ablation) | DMMGR · FVQA | top-1 accuracy: 78.6 -> 81.20 | Helped |
| add dense captioning information (ablation) | DMMGR · FVQA | top-3 accuracy: 90.6 -> 95.38 | Helped |
| use 1 reasoning step instead of 2 (ablation) | DMMGR · KRVQR | top-1 accuracy: 31.4 -> 30.5 | Hurt |
| use 3 reasoning steps instead of 2 (ablation) | DMMGR · KRVQR | top-1 accuracy: 31.4 -> 27.1 | Hurt |
| use 4 reasoning steps instead of 2 (ablation) | DMMGR · KRVQR | top-1 accuracy: 31.4 -> 26.2 | Hurt |
| replace average embedding memory with dynamic key-value memory (ablation) | DMMGR · KRVQR | top-1 accuracy: 27.3 -> 31.4 | Helped |
| replace regular key-value memory with dynamic key-value memory (ablation) | DMMGR · KRVQR | top-1 accuracy: 26.5 -> 31.4 | Helped |
| replace regular key-value memory with average embedding memory (ablation) | DMMGR · KRVQR | top-1 accuracy: 26.5 -> 27.3 | Helped |
| add knowledge-guided reasoning to the image graph module (ablation) | DMMGR · KRVQR | accuracy: 29.9 -> 31.4 | Helped |
Closest papers by meaning
- OK-VQA: A Visual Question Answering Benchmark Requiring External Knowledge (CVPR 2019 · similarity 0.92)
- TOA: Task-oriented Active VQA (NeurIPS 2023 · similarity 0.92)
- Separation of Powers: On Segregating Knowledge from Observation in LLM-enabled Knowledge-based Visual Question Answering (CVPR 2025 · similarity 0.92)
- REVIVE: Regional Visual Representation Matters in Knowledge-Based Visual Question Answering (NeurIPS 2022 · similarity 0.92)
- Prompting Large Language Models With Answer Heuristics for Knowledge-Based Visual Question Answering (CVPR 2023 · similarity 0.92)
- BOK-VQA: Bilingual outside Knowledge-Based Visual Question Answering via Graph Representation Pretraining (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.