Research map / Video / Action recognition
P3IV: Probabilistic Procedure Planning From Instructional Videos With Weak Supervision
He Zhao, Isma Hadji, Nikita Dvornik, Konstantinos G. Derpanis, Richard P. Wildes, Allan D. Jepson
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 language-based contrastive supervision to action supervision (ablation) | P3IV · CrossTask T=3 | SR: 16.90 -> 22.12 | Improved |
| Add adversarial loss to action and language supervision (ablation) | P3IV · CrossTask T=3 | SR: 22.12 -> 23.34 | Improved |
| Use Viterbi post-processing (ablation) | P3IV · NIV | SR: 20.18 -> 24.68 | Improved |
| Use a learned memory of size 128 rather than no memory (ablation) | P3IV · CrossTask T=3 | SR: 7.49 -> 23.34 | Improved |
| Increase memory size from 128 to 256 (ablation) | P3IV · CrossTask T=3 | SR: 23.34 -> 20.81 | Hurt |
| Use a probabilistic adversarial model rather than a deterministic model (ablation) | P3IV · CrossTask T=3 | KL-Div: 2.31 -> 2.11 | Improved |
| Add diversity regularization to probabilistic training (ablation) | P3IV · CrossTask T=3 | NLL: 5.07 -> 4.89 | Improved |
| Use HowTo100M-pretrained joint video-language features instead of CrossTask-trained features (ablation) | P3IV · CrossTask T=3 | SR: 10.82 -> 23.34 | Improved |
| Replace weak language supervision with strong visual supervision (ablation) | P3IV · CrossTask T=3 | SR: 23.34 -> 24.41 | Mixed |
| Increase inference samples from 150 to 2500 (ablation) | P3IV probabilistic · CrossTask T=3 | NLL: 5.03 ± 0.21 -> 4.89 ± 0.04 | Improved |
| Use a probabilistic rather than deterministic model (ablation) | P3IV · COIN | NLL: 6.37 -> 6.37 | No clear effect |
Later papers that compared with this method
3 later accepted papers report P3IV as a baseline in their results tables; 3 of them report a better number on most of the results they share with it.
- PDPP:Projected Diffusion for Procedure Planning in Instructional Videos (CVPR 2023 · better on 11 of 17 shared results)
- Skip-Plan: Procedure Planning in Instructional Videos via Condensed Action Space Learning (ICCV 2023 · better on 13 of 13 shared results)
- Event-Guided Procedure Planning from Instructional Videos with Text Supervision (ICCV 2023 · better on 9 of 9 shared results)
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Event-Guided Procedure Planning from Instructional Videos with Text Supervision (ICCV 2023 · similarity 0.91)
- RAP: Retrieval-Augmented Planner for Adaptive Procedure Planning in Instructional Videos (ECCV 2024 · similarity 0.90)
- Procedure Knowledge Decoupled Distillation Strategy for Procedure Planning in Instructional Videos (AAAI 2025 · similarity 0.89)
- Why Not Use Your Textbook? Knowledge-Enhanced Procedure Planning of Instructional Videos (CVPR 2024 · similarity 0.88)
- Masked Temporal Interpolation Diffusion for Procedure Planning in Instructional Videos (ICLR 2025 · similarity 0.88)
- SCHEMA: State CHangEs MAtter for Procedure Planning in Instructional Videos (ICLR 2024 · similarity 0.87)
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.
