Multimodality · 2021
ALIGN: Scaling Up Visual and Vision-Language Representation Learning
ALIGN showed that a dual-encoder image-text model trained with contrastive learning on a billion-scale noisy web alt-text dataset matches curated-data methods, removing the dependency on expensive human-labeled or filtered training pairs.
Editorial record
Plain-language summary
ALIGN trains separate image and text encoders to map matched image-caption pairs close together in a shared embedding space using a contrastive loss, learned over ~1.8 billion raw web image/alt-text pairs with only minimal frequency-based cleaning. By tolerating noisy supervision at scale, it removed the bottleneck of building large curated vision-language datasets and enabled strong zero-shot classification and image-text retrieval from web data alone.
Knowledge graph
Relationships
Antecedents
Parallel workEvidence: Strongly supported
CLIP: Learning Transferable Visual Models from NL Supervision
ALIGN is concurrent contrastive VL pretraining
P-302
Source record
Provenance
- Record ID
- P-302
- Record created
- 2026-07-13
- Last reviewed
- 2026-07-14
- Record version
- 2
- https://arxiv.org/abs/2102.05918
- arXiv:2102.05918
Citation caveat: Citation metadata is approximate and marked unverified in the source dataset.