I-JEPA: context-to-target image-block representation prediction

What changed

Predicted target-block embeddings from informative context blocks with a joint-embedding predictive architecture and multi-block masking, learning image features without handcrafted view-augmentation invariances.

Scientific scope

Predicted target-block embeddings from informative context blocks with a joint-embedding predictive architecture and multi-block masking, learning image features without handcrafted view-augmentation invariances. Inherits joint-embedding/teacher-student representation prediction; does not reconstruct pixels or establish a complete physical world model. No view augmentation does not mean no masking/cropping design choices.

Evidence

The linked original sources describe the stated contribution and its evidence limits. WRS has not independently rerun the experiments or certified a proof. Inherits joint-embedding/teacher-student representation prediction; does not reconstruct pixels or establish a complete physical world model. No view augmentation does not mean no masking/cropping design choices.

Contributors

• Source-listed authors and technical contributors · SCIENCE: Source-listed authors or explicitly credited technical contributors for the cited versions. Family entries retain source-specific attribution; a combined list is not one original-paper byline or equal individual work. Names alone do not establish resolved identities.

• Self-Supervised Learning from Images with a Joint-Embedding Predictive Architecture

Provisional WRS assessment, edition 2026-10-16.1. Independent expert calibration is pending. This grade does not certify correctness or assign ownership.

Full scientific record · Browse contribution grades · Request a correction