Self-supervised learning
Technique · Training · 2 items since Apr 2026
Learning representations from unlabeled data by predicting hidden or transformed parts of it.
Related
- Techniques
Items
- Sep 18
JEPA-Anything extends world models across seven domains
A domain-agnostic world-modeling framework extends joint-embedding predictive architectures with orthogonal predictive factorization, generalizing across seven domains including clinical trajectories and molecular dynamics, with a factor-nominated biological intervention validated in organoids and mice.
- Apr 14
Self-supervised tasks boost visual instruction tuning
Valeo and ETH Zurich find multimodal LLMs lean on language priors during instruction tuning, and that mixing in 3-10% visually grounded self-supervised tasks phrased as instructions consistently improves vision-centric benchmarks with no architecture changes or extra annotation.