Research map / Vision-language and multimodal / Visual question answering
Multi-Agent Embodied Question Answering in Interactive Environments
Sinan Tan, Weilai Xiang, Huaping Liu, Di Guo, Fuchun Sun
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 |
|---|---|---|---|
| replace 2D top-down memory with 3D multi-layer semantic memory (GT segmentation) | IQA baseline · IQuADv1 Existence | accuracy: 86.56% -> 98.75% | Improved |
| replace 2D top-down memory with 3D memory (GT segmentation) | IQA baseline · IQuADv1 Counting | accuracy: 35.31% -> 88.28 % | Improved |
| replace 2D memory with 3D memory (GT segmentation) | IQA baseline · IQuADv1 Containing | accuracy: 70.94% -> 91.88% | Improved |
| use 3D-memory framework with GT segmentation instead of IQA with GT detection | IQA baseline · IQuADv1 Existence | episode length: 679.70 -> 166.31 | Improved |
| use predicted Mask R-CNN segmentation instead of GT segmentation | Ours · IQuADv1 Existence | accuracy: 98.75% -> 79.53% | Hurt |
| use predicted segmentation instead of GT segmentation | Ours · IQuADv1 Counting | accuracy: 88.28 % -> 45.62% | Hurt |
| use predicted segmentation instead of GT segmentation | Ours · IQuADv1 Containing | accuracy: 91.88% -> 77.50% | Hurt |
| use predicted segmentation framework vs IQA with GT detection | Ours (Pred. Segmentation) · IQuADv1 Existence | accuracy: 86.56% -> 79.53% | Hurt |
| increase number of agents from 1 to 4 | Ours · IQuADv1 overall | accuracy: 67.55% -> 66.20% | No clear effect |
| increase number of agents from 1 to 2 | Ours · IQuADv1 overall | episode length: 195.22 -> 112.75 | Improved |
| increase number of agents from 1 to 4 | Ours · IQuADv1 overall | episode length: 195.22 -> 71.83 | Improved |
| increase number of agents from 2 to 4 | Ours · IQuADv1 overall | speedup as % of ideal: 87% -> 68% | Hurt |
Closest papers by meaning
- OpenEQA: Embodied Question Answering in the Era of Foundation Models (CVPR 2024 · similarity 0.89)
- Multi-modal Situated Reasoning in 3D Scenes (NeurIPS 2024 · similarity 0.88)
- Exemplar2VQA: A Scalable Exemplar-Driven Visual Question Answering Generation Framework via Multi-Agent Coding (NeurIPS 2026 · similarity 0.88)
- Seeing, Listening, Remembering, and Reasoning: A Multimodal Agent with Long-Term Memory (ICLR 2026 · similarity 0.87)
- Refine and Align: Confidence Calibration Through Multi-Agent Interaction in VQA (AAAI 2026 · similarity 0.85)
- IndustryEQA: Pushing the Frontiers of Embodied Question Answering in Industrial Scenarios (NeurIPS 2025 · similarity 0.85)
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.
