Research map / Vision-language and multimodal / Robot action and navigation
A4-Agent: An Agentic Framework for Zero-Shot Affordance Reasoning
Zixin Zhang, Kanghao Chen, Hanqing Wang, Hongfei Zhang, Harold Haodong Chen, Chenfei Liao, Litao Guo, Yinchuan Li, Yingcong Chen
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 |
|---|---|---|---|
| add think-with-imagination (Dreamer) (ablation) | A4-Agent (Qwen-2.5-VL-7B) · RAGNet-3DOI | gIoU: 58.48 -> 63.02 | Improved |
| add think-with-imagination (Dreamer) (ablation) | A4-Agent (Qwen-2.5-VL-7B) · RAGNet-3DOI | cIoU: 49.26 -> 49.87 | No clear effect |
| add think-with-imagination (Dreamer) (ablation) | A4-Agent (GPT-4o) · RAGNet-3DOI | cIoU: 54.43 -> 58.30 | Improved |
| add think-with-imagination (Dreamer) (ablation) | A4-Agent (GPT-4o) · RAGNet-3DOI | gIoU: 62.30 -> 63.94 | Improved |
| add Thinker (VLM reasoning) before Spotter (ablation) | A4-Agent · RAGNet-3DOI | gIoU: 45.91 -> 62.30 | Improved |
| replace Qwen-2.5-VL-7B reasoning backbone with GPT-4o (ablation) | A4-Agent · RAGNet-3DOI | gIoU: 58.48 -> 62.30 | Improved |
| shrink segmentation backbone from SAM2-Large to SAM2-Tiny (ablation) | A4-Agent · RAGNet-3DOI | gIoU: 58.48 -> 56.32 | Hurt |
| use decoupled zero-shot agent versus trained end-to-end Affordance-R1 | A4-Agent · ReasonAff | gIoU: 67.41 -> 70.52 | Improved |
| zero-shot agent versus supervised AffordanceVLM | A4-Agent · RAGNet-HANDAL-easy | gIoU: 58.3 -> 61.1 | Improved |
| zero-shot agent versus Affordance-R1 on P50-95 | A4-Agent · ReasonAff | P50-95: 55.22 -> 55.22 | No clear effect |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- Affordance-R1 — Affordance-R1: Reinforcement Learning for Generalizable Affordance Reasoning in Multimodal Large Language Models (AAAI 2026) · better on 5 of 5 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- AffordBot: 3D Fine-grained Embodied Reasoning via Multimodal Large Language Models (NeurIPS 2025 · similarity 0.89)
- Affordance-R1: Reinforcement Learning for Generalizable Affordance Reasoning in Multimodal Large Language Models (AAAI 2026 · similarity 0.89)
- IAAO: Interactive Affordance Learning for Articulated Objects in 3D Environments (CVPR 2025 · similarity 0.88)
- Probing and Bridging Geometry-Interaction Cues for Affordance Reasoning in Vision Foundation Models (CVPR 2026 · similarity 0.87)
- EmbodiedBench: Comprehensive Benchmarking Multi-modal Large Language Models for Vision-Driven Embodied Agents (ICML 2025 · similarity 0.87)
- SeqAfford: Sequential 3D Affordance Reasoning via Multimodal Large Language Model (CVPR 2025 · similarity 0.87)
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.
