Research map / Detection and segmentation / Open-set and few-shot detection
Unsupervised Learning of Object Keypoints for Perception and Control
Tejas D Kulkarni, Ankush Gupta, Catalin Ionescu, Sebastian Borgeaud, Malcolm Reynolds, Andrew Zisserman, Volodymyr Mnih
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 |
|---|---|---|---|
| Use feature transport between frames with a keypoint bottleneck | Transporter · Atari ALE and Manipulator tracking | long-term keypoint tracking | Improved |
| Use Transporter keypoints and corresponding image features as policy inputs | KeyQN · Atari ALE, 5 games | mean game score | Improved |
| Explore by controlling keypoint coordinates with temporally extended options | Transporter keypoint options · Montezuma's Revenge, Ms. Pac-Man, Seaquest | episodic return | Improved |
| Transport learned features rather than pixels (ablation) | Transporter · Manipulator tabletop toys | pixel reconstruction error | Improved |
| Compare Transporter with Zhang et al. on Ms. Pac-Man tracking | Transporter · Ms. Pac-Man | long-term keypoint tracking | Mixed |
Closest papers by meaning
- Unsupervised learning of object structure and dynamics from videos (NeurIPS 2019 · similarity 0.88)
- 3D Implicit Transporter for Temporally Consistent Keypoint Discovery (ICCV 2023 · similarity 0.88)
- Entropy-driven Unsupervised Keypoint Representation Learning in Videos (ICML 2023 · similarity 0.88)
- Unsupervised Learning of Visual 3D Keypoints for Control (ICML 2021 · similarity 0.87)
- Learning Features by Watching Objects Move (CVPR 2017 · similarity 0.87)
- Control-oriented Clustering of Visual Latent Representation (ICLR 2025 · similarity 0.85)
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.
