Research map / Graph learning / Graph self-supervised learning
Space-Time Correspondence as a Contrastive Random Walk
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 test-time self-supervised adaptation (ablation) | Space-Time Correspondence as a Contrastive Random Walk · DAVIS 2017 val | J&F mean: 67.6 -> 68.3 | Helped |
| add moderate edge dropout (ablation) | Space-Time Correspondence as a Contrastive Random Walk · DAVIS 2017 val | object correspondence performance | Helped |
| train on longer video sequences (ablation) | Space-Time Correspondence as a Contrastive Random Walk · DAVIS 2017 val | object segmentation performance | Helped |
| train using temporally changing video rather than the same image at each step (ablation) | Space-Time Correspondence as a Contrastive Random Walk · DAVIS 2017 val | J&F mean: 57.5 -> 65.9; 67.5; 67.6 | Helped |
| increase training frame rate to 30 (ablation) | Space-Time Correspondence as a Contrastive Random Walk · DAVIS 2017 val | J&F mean: 67.6 -> 62.3 | Hurt |
| add more temporal context at evaluation (ablation) | Space-Time Correspondence as a Contrastive Random Walk · VIP | Parts score: 36.0 -> 38.6 | Helped |
| add more temporal context at evaluation (ablation) | Space-Time Correspondence as a Contrastive Random Walk · JHMDB | Pose score (first reported column): 59.0 -> 59.3 | Helped |
| add more temporal context at evaluation (ablation) | Space-Time Correspondence as a Contrastive Random Walk · JHMDB | Pose score (second reported column): 83.2 -> 84.9 | Helped |
| restrict source nodes using spatial locality (ablation) | ImageNet label propagation baseline · DAVIS 2017 val | J&F: +4% J&F | Helped |
| select top-k neighbors over all source nodes rather than separately per frame (ablation) | ImageNet label propagation baseline · DAVIS 2017 val | J&F: +3% J&F | Helped |
Closest papers by meaning
- Exposing the Self-Supervised Space-Time Correspondence Learning via Graph Kernels (AAAI 2023 · similarity 0.93)
- Contrastive Learning for Space-Time Correspondence via Self-Cycle Consistency (CVPR 2022 · similarity 0.92)
- Self-Supervised Cross-View Correspondence with Predictive Cycle Consistency (CVPR 2025 · similarity 0.88)
- Self-Supervised Spatial Correspondence Across Modalities (CVPR 2025 · similarity 0.88)
- Emergent Temporal Correspondences from Video Diffusion Transformers (NeurIPS 2025 · similarity 0.87)
- Video Diffusion Models Excel at Tracking Similar-Looking Objects Without Supervision (NeurIPS 2025 · 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.