Research map / 3D vision / Human pose estimation
Adversarial Generative Grammars for Human Activity Prediction
AJ Piergiovanni, Anelia Angelova, Alexander Toshev, Michael S. Ryoo
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 branching to grammar-only model (ablation) | Grammar only · Charades (45 sec) | accuracy: 3.8 -> 8.5 | Improved |
| add branching to grammar-only model (ablation) | Grammar only · Charades (1 sec) | accuracy: 12.2 -> 15.7 | Improved |
| add branching to adversarial grammar (ablation) | AGG · Charades (45 sec) | accuracy: 5.5 -> 8.8 | Improved |
| add branching to adversarial grammar (ablation) | AGG · Charades (1 sec) | accuracy: 14.2 -> 15.9 | Improved |
| add adversarial training to grammar (ablation) | Grammar only · Charades (45 sec, single output) | accuracy: 8.5 -> 8.8 | Improved |
| add adversarial training to grammar (ablation) | Grammar only · Charades (10 sec) | accuracy: 11.2 -> 10.7 | Hurt |
| add adversarial training to grammar (ablation) | Grammar only · MultiTHUMOS (45 sec, single output) | mAP: 8.5 -> 11.4 | Improved |
| add adversarial training to grammar (ablation) | Grammar only · 50 Salads (20% observed, 50% predicted) | accuracy: 18.7 -> 20.9 | Improved |
| add adversarial training to grammar (ablation) | Grammar only · 50 Salads (30% observed, 20% predicted) | accuracy: 30.1 -> 40.1 | Improved |
| generate 10 futures and take best instead of single output (ablation) | AGG-single · MultiTHUMOS (1 sec) | mAP: 19.3 -> 22.0 | Improved |
| generate 10 futures and take best instead of single output (ablation) | AGG-single · MultiTHUMOS (45 sec) | mAP: 11.4 -> 11.4 | No clear effect |
| generate 10 futures and take best instead of single output (ablation) | AGG-single · Charades (45 sec) | accuracy: 8.8 -> 9.8 | Improved |
Closest papers by meaning
- Predicting Human Activities Using Stochastic Grammar (ICCV 2017 · similarity 0.88)
- Generative Hybrid Representations for Activity Forecasting With No-Regret Learning (CVPR 2020 · similarity 0.86)
- Predicting the Future: A Jointly Learnt Model for Action Anticipation (ICCV 2019 · similarity 0.86)
- On Diverse Asynchronous Activity Anticipation (ECCV 2020 · similarity 0.85)
- FutureHuman3D: Forecasting Complex Long-Term 3D Human Behavior from Video Observations (CVPR 2024 · similarity 0.85)
- Social GAN: Socially Acceptable Trajectories With Generative Adversarial Networks (CVPR 2018 · 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.
