TC-JEPA conditions masked feature prediction on text captions via sparse cross-attention to produce more semantically rich visual representations and outperforms contrastive methods on fine-grained tasks.
Self-supervised learning from images with a joint-embedding predictive architecture
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Text-Conditional JEPA for Learning Semantically Rich Visual Representations
TC-JEPA conditions masked feature prediction on text captions via sparse cross-attention to produce more semantically rich visual representations and outperforms contrastive methods on fine-grained tasks.