Summary of Revisiting Feature Prediction For Learning Visual Representations From Video, by Adrien Bardes et al.
Revisiting Feature Prediction for Learning Visual Representations from Video
by Adrien Bardes, Quentin Garrido, Jean Ponce, Xinlei Chen, Michael Rabbat, Yann LeCun, Mahmoud Assran, Nicolas Ballas
First submitted to arxiv on: 15 Feb 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary The paper introduces V-JEPA, a collection of vision models trained solely using feature prediction from video data, without relying on pre-trained image encoders or other forms of supervision. The models are trained on 2 million videos and evaluated on various downstream tasks, such as image and video classification. The results show that these models learn versatile visual representations that perform well on both motion- and appearance-based tasks, without requiring adaptation of the model’s parameters. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper explores a new approach to learning from video data by predicting features, rather than relying on pre-trained models or other sources of supervision. The authors introduce V-JEPA, a collection of vision models that are trained solely using feature prediction from videos. These models are tested on various tasks and shown to be effective, without requiring any additional training or fine-tuning. |
Keywords
* Artificial intelligence * Classification * Fine tuning
