Research map / Probabilistic and scientific modeling / Dynamical systems
Generative Hybrid Representations for Activity Forecasting With No-Regret Learning
Jiaqi Guan, Ye Yuan, Kris M. Kitani, Nicholas Rhinehart
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 |
|---|---|---|---|
| Condition action forecasts on generated trajectories through joint training (ablation) | Hybrid activity forecasting · EPIC-KITCHENS test | action F1: 34.14 -> 37.90 | Improved |
| Add the reverse cross entropy terms to forward-cross-entropy training (ablation) | Joint hybrid activity forecasting · EPIC-KITCHENS test | trajectory meanMSD: 1.446 ± 0.087 -> 0.971 ± 0.078 | Improved |
| Add the reverse cross entropy terms to forward-cross-entropy training (ablation) | Joint hybrid activity forecasting · EPIC-KITCHENS test | action F1: 36.98 ± 0.30 -> 37.90 ± 0.09 | Improved |
| Add the reverse cross entropy terms to forward-cross-entropy training (ablation) | Joint hybrid activity forecasting · EPIC-KITCHENS test | trajectory forward cross entropy: -298.92 -> -298.47 | No clear effect |
| Use joint training rather than separate trajectory and action training (ablation) | Hybrid activity forecasting · EPIC-KITCHENS test | trajectory minMSD: 0.286 ± 0.007 -> 0.293 ± 0.004 | Mixed |
| Fine-tune the last linear layer online | Ours (J) · Train → Test streaming data | action F1: 37.90 ± 0.09 -> 38.11 ± 0.10 | Improved |
| Fine-tune the last linear layer online | Ours (J) · Test → Train streaming data | action F1: 20.49 ± 0.02 -> 22.40 ± 0.05 | Improved |
| Use a generative hybrid forecaster rather than CVAE | EPIC-KITCHENS activity forecasting · EPIC-KITCHENS test | trajectory minMSD: 0.319 ± 0.008 -> 0.293 ± 0.004 | Improved |
| Use a generative hybrid forecaster rather than MRMC | EPIC-KITCHENS activity forecasting · EPIC-KITCHENS test | trajectory meanMSD: 0.392 -> 0.971 ± 0.078 | Hurt |
Closest papers by meaning
- A Hybrid Egocentric Activity Anticipation Framework via Memory-Augmented Recurrent and One-Shot Representation Forecasting (CVPR 2022 · similarity 0.88)
- First-Person Activity Forecasting With Online Inverse Reinforcement Learning (ICCV 2017 · similarity 0.87)
- FutureHuman3D: Forecasting Complex Long-Term 3D Human Behavior from Video Observations (CVPR 2024 · similarity 0.86)
- Adversarial Generative Grammars for Human Activity Prediction (ECCV 2020 · similarity 0.86)
- First-Person Activity Forecasting from Video with Online Inverse Reinforcement Learning (TPAMI 2018 · similarity 0.86)
- Predicting the Future: A Jointly Learnt Model for Action Anticipation (ICCV 2019 · 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.
