AWhat changed
Used teacher-student self-distillation on masked patch tokens and global image tokens so the visual tokenizer learns jointly with masked-image pretraining.
Scientific scope
Used teacher-student self-distillation on masked patch tokens and global image tokens so the visual tokenizer learns jointly with masked-image pretraining. Masked modeling, DINO-style distillation and teacher networks are inherited; joint online tokenization/patch supervision is the scoped change. Classification and dense-task transfer are empirical, not a universal semantic tokenizer. Original2021 preprint precedes ICLR2022.
Evidence
The linked original sources describe the stated contribution and its evidence limits. WRS has not independently rerun the experiments or certified a proof. Masked modeling, DINO-style distillation and teacher networks are inherited; joint online tokenization/patch supervision is the scoped change. Classification and dense-task transfer are empirical, not a universal semantic tokenizer. Original2021 preprint precedes ICLR2022.
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.
- iBOT: Image BERT Pre-Training with Online Tokenizer
- Zhou et al.: iBOT masked image modeling with an online tokenizer — author/institution implementation
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