Research map / Vision-language and multimodal / Robot action and navigation
Token-Based Affordance Grounding with Large Vision-Language Models
Seung Il Lee, Qinqian Lei, Daguang Xu, Dong Yang, Robby T. Tan, Yixin Chen, Bo Wang
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 |
|---|---|---|---|
| aggregate attention over all layers and heads instead of layer 25 only (ablation) | TokAG (2B) · AGD20K Set 2 (Unseen) | NSS: 1.471 -> 1.514 | Improved |
| aggregate over all layers and heads instead of layer 16 only (ablation) | TokAG (2B) · AGD20K Set 2 (Unseen) | NSS: 1.466 -> 1.514 | Improved |
| aggregate over all layers and heads instead of head 5 only (ablation) | TokAG (2B) · AGD20K Set 2 (Unseen) | NSS: 1.510 -> 1.514 | Improved |
| select spatial-aware top-1 token instead of averaging all tokens (ablation) | TokAG (2B) · AGD20K Set 2 (Unseen) | NSS: 1.457 -> 1.514 | Improved |
| select spatial-aware top-1 token instead of last output token (ablation) | TokAG (2B) · AGD20K Set 2 (Unseen) | NSS: 1.116 -> 1.514 | Improved |
| select spatial-aware top-1 token instead of first output token (ablation) | TokAG (2B) · AGD20K Set 2 (Unseen) | NSS: 1.493 -> 1.514 | Improved |
| select spatial-aware top-1 token instead of nouns only (ablation) | TokAG (2B) · AGD20K Set 2 (Unseen) | NSS: 1.461 -> 1.514 | Improved |
| use top-1 token instead of top-5 tokens (ablation) | TokAG (2B) · AGD20K Set 2 (Unseen) | SIM: SIM 0.444 vs 0.455; KLD 1.084 vs 1.060; NSS 1.482 vs 1.514 | Mixed |
| replace CLIPSeg object mask with SAM3 (ablation) | TokAG (2B) · AGD20K Set 2 (Unseen) | NSS: 1.514 -> 1.416 | Hurt |
| replace Qwen3-VL (2B) backbone with InternVL3 (2B) (ablation) | TokAG · AGD20K Set 2 (Unseen) | NSS: 1.514 -> 1.323 | Hurt |
| scale LVLM from 2B to 32B | TokAG · AGD20K Set 2 (Unseen) | NSS: 1.514 -> 1.549 | Improved |
| scale LVLM from 2B to 32B | TokAG · HICO-IIF (Seen) | NSS: 1.600 -> 1.655 | Improved |
Closest papers by meaning
- Weakly-Supervised Affordance Grounding Guided by Part-Level Semantic Priors (ICLR 2025 · similarity 0.90)
- Weakly Supervised Multimodal Affordance Grounding for Egocentric Images (AAAI 2024 · similarity 0.89)
- Reasoning Mamba: Hypergraph-Guided Region Relation Calculating for Weakly Supervised Affordance Grounding (CVPR 2025 · similarity 0.89)
- Grounding 3D Object Affordance with Language Instructions, Visual Observations and Interactions (CVPR 2025 · similarity 0.89)
- Your Large Vision-Language Model Only Needs A Few Attention Heads For Visual Grounding (CVPR 2025 · similarity 0.88)
- VideoGEM: Training-free Action Grounding in Videos (CVPR 2025 · similarity 0.88)
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.
