Research map / Label-efficient and robust learning / Recognition and long-tail
Commonly Uncommon: Semantic Sparsity in Situation Recognition
Mark Yatskar, Vicente Ordonez, Luke Zettlemoyer, Ali Farhadi
Check 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 tensor composition with image regression potentials (ablation) | Image Regression CRF · imSitu development | top-5 verb accuracy: 58.64 -> 59.92 | Improved |
| Add semantic augmentation to the image regression baseline (ablation) | Image Regression CRF · imSitu rare development | top-5 verb accuracy: 44.00 -> 43.08 | Hurt |
| Add semantic augmentation to tensor composition with regression (ablation) | Tensor composition + reg · imSitu development | top-5 verb accuracy: 59.92 -> 61.75 | Improved |
| Add self-training after semantic augmentation (ablation) | Tensor composition + reg + SA · imSitu development | top-5 verb accuracy: 61.75 -> 62.21 | Improved |
| Add self-training after semantic augmentation (ablation) | Tensor composition + reg + SA · imSitu development | value accuracy given ground truth verbs: 35.08 -> 34.82 | Hurt |
| Add semantic augmentation to tensor composition with regression (ablation) | Tensor composition + reg · imSitu rare development | top-5 verb accuracy: 45.14 -> 44.82 | Improved |
| Add self-training after semantic augmentation (ablation) | Tensor composition + reg + SA · imSitu rare development | top-5 verb accuracy: 44.82 -> 45.94 | Improved |
| Combine tensor composition, image regression, semantic augmentation, and self-training | Image Regression CRF · imSitu test | top-5 verb accuracy: 58.88 -> 62.59 | Improved |
Closest papers by meaning
- Situation Recognition: Visual Semantic Role Labeling for Image Understanding (CVPR 2016 · similarity 0.89)
- Grounded Situation Recognition (ECCV 2020 · similarity 0.88)
- Grounded Video Situation Recognition (NeurIPS 2022 · similarity 0.87)
- ORIC: Benchmarking Object Recognition under Contextual Incongruity in Large Vision-Language Models (CVPR 2026 · similarity 0.85)
- DoGA: Enhancing Grounded Object Detection via Grouped Pre-Training with Attributes (AAAI 2025 · similarity 0.85)
- Outlier Weighed Layerwise Sparsity (OWL): A Missing Secret Sauce for Pruning LLMs to High Sparsity (ICML 2024 · similarity 0.84)
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.
