Research map / Vision-language and multimodal / Human and embodied modeling
Grounded Situation Recognition
Sarah Pratt, Mark Yatskar, Luca Weihs, Ali Farhadi, Aniruddha Kembhavi
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 |
|---|---|---|---|
| train situation recognition and grounding jointly (JSL) instead of independently (ISL) | ISL · SWiG dev (top-1 verb) | grounded-value-all: 7.64 -> 10.16 | Improved |
| train situation recognition and grounding jointly (JSL) instead of independently (ISL) | ISL · SWiG dev (top-1 verb) | grounded-value: 22.47 -> 25.03 | Improved |
| train jointly instead of independently | ISL · SWiG dev (ground-truth verb) | grounded-value: 52.92 -> 57.50 | Improved |
| train jointly instead of independently | ISL · SWiG dev (ground-truth verb) | grounded-value-all: 15.00 -> 19.29 | Improved |
| train jointly instead of independently | ISL · SWiG dev (top-5 verbs) | grounded-value-all: 11.66 -> 15.07 | Improved |
| train jointly instead of independently | ISL · SWiG dev (top-1 verb) | verb accuracy: 38.83 -> 39.60 | Improved |
| train jointly instead of independently | ISL · SWiG dev (top-1 verb) | value: 30.47 -> 31.18 | Improved |
| train jointly instead of independently | ISL · SWiG dev (ground-truth verb) | value: 72.77 -> 73.53 | Improved |
| train jointly instead of independently | ISL · SWiG test (ground-truth verb) | grounded-value-all: 14.58 -> 18.45 | Improved |
| train jointly instead of independently | ISL · SWiG test (ground-truth verb) | grounded-value: 52.19 -> 56.57 | Improved |
| upgrade RNN w/o Fusion (reimplementation, new backbone, label smoothing, tuning) | RNN w/o Fusion · SWiG dev (top-1 verb) | verb accuracy: 35.35 -> 38.83 | Improved |
| upgrade RNN w/o Fusion (reimplementation, new backbone, label smoothing, tuning) | RNN w/o Fusion · SWiG dev (top-1 verb) | value-all: 15.77 -> 18.23 | Improved |
Closest papers by meaning
- Grounded Video Situation Recognition (NeurIPS 2022 · similarity 0.90)
- Situation Recognition: Visual Semantic Role Labeling for Image Understanding (CVPR 2016 · similarity 0.90)
- Comprehensive Visual Grounding for Video Description (AAAI 2024 · similarity 0.87)
- GroundVLP: Harnessing Zero-Shot Visual Grounding from Vision-Language Pre-training and Open-Vocabulary Object Detection (AAAI 2024 · similarity 0.86)
- GroundingME: Exposing the Visual Grounding Gap in MLLMs through Multi-Dimensional Evaluation (CVPR 2026 · similarity 0.86)
- Connecting the Dots: Training-Free Visual Grounding via Agentic Reasoning (AAAI 2026 · 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.
