Research map / Vision-language and multimodal / Visual question answering
Embodied Question Answering
Abhishek Das, Samyak Datta, Georgia Gkioxari, Stefan Lee, Devi Parikh, Dhruv Batra
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 |
|---|---|---|---|
| use an ACT planner-controller instead of a standard LSTM navigator (ablation) | EmbodiedQA navigation · EQA v1 test, T_-10 | d_T (m): 1.13 -> 0.46 | Improved |
| use an ACT planner-controller instead of a standard LSTM navigator (ablation) | EmbodiedQA navigation · EQA v1 test, T_-30 | d_T (m): 2.23 -> 1.50 | Improved |
| use an ACT planner-controller instead of a standard LSTM navigator (ablation) | EmbodiedQA navigation · EQA v1 test, T_-50 | d_T (m): 2.89 -> 2.74 | Improved |
| add question encoding to a reactive navigator (ablation) | Reactive navigation · EQA v1 test, T_-10 | d_T (m): 2.09 -> 1.58 | Improved |
| add question encoding to an LSTM navigator (ablation) | LSTM navigation · EQA v1 test, T_-10 | d_T (m): 1.75 -> 1.13 | Improved |
| add memory to a reactive navigator (ablation) | Question-agnostic navigation · EQA v1 test, T_-10 | d_T (m): 2.09 -> 1.75 | Improved |
| add memory to a question-conditioned reactive navigator (ablation) | Question-conditioned navigation · EQA v1 test, T_-10 | d_T (m): 1.58 -> 1.13 | Improved |
| fine-tune navigation with reinforcement learning and target-distance shaping (ablation) | ACT+Q · EQA v1 test, T_-30 | mean rank: 3.13 -> 2.99 | Improved |
| fine-tune navigation with reinforcement learning and target-distance shaping (ablation) | ACT+Q · EQA v1 test, T_-10 | mean rank: 3.09 -> 3.13 | Hurt |
| fine-tune navigation with reinforcement learning and target-distance shaping (ablation) | ACT+Q · EQA v1 test, T_-50 | mean rank: 3.25 -> 3.22 | Improved |
| pretrain perception heads jointly rather than independently (ablation) | CNN segmentation · Perception validation set | mean IOU: 0.163 -> 0.166 | Improved |
| use human navigation views rather than shortest-path views | Question answering module · Human navigation | mean rank: 3.26 -> 3.51 | Hurt |
Closest papers by meaning
- OpenEQA: Embodied Question Answering in the Era of Foundation Models (CVPR 2024 · similarity 0.92)
- Multi-modal Situated Reasoning in 3D Scenes (NeurIPS 2024 · similarity 0.88)
- Embodied Scene Understanding for Vision Language Models via MetaVQA (CVPR 2025 · similarity 0.88)
- IndustryEQA: Pushing the Frontiers of Embodied Question Answering in Industrial Scenarios (NeurIPS 2025 · similarity 0.87)
- Situational Awareness Matters in 3D Vision Language Reasoning (CVPR 2024 · similarity 0.87)
- ECBench: Can Multi-modal Foundation Models Understand the Egocentric World? A Holistic Embodied Cognition Benchmark (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.
