Research map / Vision-language and multimodal / Visual question answering
AssistQ: Affordance-Centric Question-Driven Task Completion for Egocentric Assistant
Benita Wong, Joya Chen, You Wu, Stan Weixian Lei, Dongxing Mao, Difei Gao, Mike Zheng Shou
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 reverse-mask button encoding only instead of no mask (ablation) | Q2A · AssistQ | Recall@1: 17.9 -> 30.2 | Improved |
| add mask to reverse-mask button encoding (double-mask) (ablation) | Q2A · AssistQ | Recall@1: 30.2 -> 21.4 | Hurt |
| add video to script input (ablation) | Q2A · AssistQ | Recall@1: 24.6 -> 30.2 | Improved |
| add script to video input (ablation) | Q2A · AssistQ | Recall@1: 29.4 -> 30.2 | Improved |
| add video and script versus neither (ablation) | Q2A · AssistQ | Recall@1: 23.4 -> 30.2 | Improved |
| add QA->S attention to context grounding (ablation) | Q2A · AssistQ | Recall@1: 19.8 -> 28.6 | Improved |
| add S->V attention to context grounding (ablation) | Q2A · AssistQ | Recall@1: 19.8 -> 22.6 | Improved |
| add transfer attention QA->S->V on top of QA->S and S->V (ablation) | Q2A · AssistQ | Recall@1: 29.0 -> 30.2 | Improved |
| add history of steps to GRU steps network (ablation) | Q2A · AssistQ | Recall@1: 25.0 -> 30.2 | Improved |
| add history of steps to MLP steps network (ablation) | Q2A · AssistQ | Recall@3: 21.4 -> 21.8 Recall@1; 60.7 -> 56.8 Recall@3 | Mixed |
| use GRU instead of MLP as steps network with history (ablation) | Q2A · AssistQ | Recall@1: 21.8 -> 30.2 | Improved |
| use Q2A instead of PACMAN | PACMAN · AssistQ | Recall@1: 24.6 -> 30.2 | Improved |
Closest papers by meaning
- Grounded Question-Answering in Long Egocentric Videos (CVPR 2024 · similarity 0.86)
- LifeEval: A Multimodal Benchmark for Assistive AI in Egocentric Daily Life Tasks (CVPR 2026 · similarity 0.86)
- In the Eye of MLLM: Benchmarking Egocentric Video Intent Understanding with Gaze-Guided Prompting (NeurIPS 2025 · similarity 0.86)
- EgoThink: Evaluating First-Person Perspective Thinking Capability of Vision-Language Models (CVPR 2024 · similarity 0.84)
- OpenEQA: Embodied Question Answering in the Era of Foundation Models (CVPR 2024 · similarity 0.84)
- ECBench: Can Multi-modal Foundation Models Understand the Egocentric World? A Holistic Embodied Cognition Benchmark (CVPR 2025 · similarity 0.83)
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.
