negative caption
VLC: Extending Vision-Language Compositionality Evaluation with Text-to-Image Retrieval
Compositionality is still a challenging problem.
- Europe > Switzerland > Zürich > Zürich (0.14)
- Europe > Spain > Basque Country (0.04)
- Overview (0.68)
- Research Report > New Finding (0.46)
- Information Technology (0.68)
- Government (0.46)
DesCo: Learning Object Recognition with Rich Language Descriptions
Recent development in vision-language approaches has instigated a paradigm shift in learning visual recognition models from language supervision. These approaches align objects with language queries (e.g. "a photo of a cat") and thus improve the models' adaptability to novel objects and domains.
- Transportation (0.93)
- Government > Regional Government (0.46)
- Europe > Switzerland > Zürich > Zürich (0.14)
- North America > United States > Arizona (0.04)
- North America > United States > Maryland > Baltimore County (0.04)
- (2 more...)
- Research Report > Experimental Study (0.93)
- Research Report > New Finding (0.93)
- Europe > Switzerland > Zürich > Zürich (0.14)
- Europe > Croatia > Dubrovnik-Neretva County > Dubrovnik (0.04)
- Asia > Middle East > UAE (0.04)
- (11 more...)
- Government (1.00)
- Law (0.92)
- Information Technology > Services (0.68)
A Limitations and societal impact A.1 Limitations B
VLC is a benchmark for Bidirectional Vision-Language Composition-ality evaluation. Each instance consists of two images and two captions.
VLC: Extending Vision-Language Compositionality Evaluation with Text-to-Image Retrieval
Compositionality is still a challenging problem.
- Europe > Switzerland > Zürich > Zürich (0.14)
- Europe > Spain > Basque Country (0.04)
- Overview (0.68)
- Research Report > New Finding (0.46)
- Information Technology (0.68)
- Government (0.46)
TripletCLIP: Improving Compositional Reasoning of CLIP via Synthetic Vision-Language Negatives Maitreya Patel
Contrastive Language-Image Pretraining (CLIP) models maximize the mutual information between textual and visual modalities to learn representations.
- Europe > Switzerland > Zürich > Zürich (0.14)
- North America > United States > Arizona (0.04)
- North America > United States > Maryland > Baltimore County (0.04)
- (2 more...)
- Research Report > Experimental Study (0.93)
- Research Report > New Finding (0.93)
- Europe > Switzerland > Zürich > Zürich (0.14)
- Europe > Croatia > Dubrovnik-Neretva County > Dubrovnik (0.04)
- Asia > Middle East > UAE (0.04)
- (11 more...)
- Government (1.00)
- Law (0.92)
- Information Technology > Services (0.68)
DesCo: Learning Object Recognition with Rich Language Descriptions
Recent development in vision-language approaches has instigated a paradigm shift in learning visual recognition models from language supervision. These approaches align objects with language queries (e.g. "a photo of a cat") and thus improve the models' adaptability to novel objects and domains.
- Transportation (0.93)
- Government > Regional Government (0.46)