Research
Temporal Self-Distillation: Learning Visual State Tracking in Videos Without Supervision
arXiv:2609.04203v1 Announce Type: new Abstract: We introduce S^3T (Self-Supervised Self-Distillation over Time), which, to the best of our knowledge, is the first fully self-contained framework for co
arXiv:2609.04203v1 Announce Type: new Abstract: We introduce S^3T (Self-Supervised Self-Distillation over Time), which, to the best of our knowledge, is the first fully self-contained framework for continuous video state tracking. Our method treats temporal sampling density as privileged information, based on the hypothesis that a denser view of the same clip recovers the running state more accurately. This view serves as the teacher, while a sparse-view student with the same weights learns to match its next-token distribution. The model generates its own target, so training requires no labels, separate teacher, or reward signal, and adds no inference cost. On LLaVA-OneVision-2-8B, S^3T improves VSTAT accuracy by +1.74 as a single model, +2.38 with souping, and +2.70 with additional vision-encoder adaptation, while prior self-evolving methods leave state tracking largely unchanged. The capability learned from unlabeled synthetic clips transfers to real videos, improving performance by +7.95 on VSTAT-YouTube state-tracking questions and +4.50 on MVBench Action Count.
Related
- Contrastive Learning under Noisy Temporal Self-Supervision for Colonoscopy Videos
- Clip-level Uncertainty and Temporal-aware Active Learning for End-to-End Multi-Object Tracking
- Three-Stream Temporal-Shift Attention Network Based on Self-Knowledge Distillation for Micro-Expression Recognition
Source: arXiv cs.CV | 2026-09-04