AI Paper Atlas

Research map / Vision-language and multimodal / CLIP and zero-shot

ASIF: Coupled Data Turns Unimodal Models to Multimodal without Training

Antonio Norelli, Marco Fumero, Valentino Maiorca, Luca Moschella, Emanuele Rodolà, Francesco Locatello

NeurIPS 2023 · Vision-language and multimodal › CLIP and zero-shot · approach: Recaptioning and text-supervised pretraining · 7 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use smaller image and text encoders (DeiT-tiny) (ablation)ASIF · ImageNet zero-shotaccuracy: 60.9 -> 46.5Hurt
use DeiT-small with smaller sentence encoder (ablation)ASIF · ImageNet zero-shotaccuracy: 60.9 -> 59.3Hurt
pretrain image encoder on ImageNet21k (ViT-b16) instead of ImageNet1k (DeiT-base) (ablation)ASIF · CIFAR100 zero-shotaccuracy: 50.2 -> 63.3Improved
pretrain image encoder on ImageNet21k (ViT-b16) instead of ImageNet1k (DeiT-base) (ablation)ASIF · ImageNet zero-shotaccuracy: 60.9 -> 55.4Hurt
unsupervised DINO visual encoder instead of supervised DeiT (ablation)ASIF · ImageNet zero-shotaccuracy: 60.9 -> 53.0Hurt
add 100 EuroSAT image-caption pairs to the multimodal dataset (ablation)ASIF · EuroSATaccuracy: 29.4 -> 82.5Improved
tune k and p on PETS instead of standard k=800, p=8 (ablation)ASIF · Petsaccuracy: 71.5 -> 72.3Improved
tune k and p on CIFAR100 instead of standard k=800, p=8 (ablation)ASIF · CIFAR100accuracy: 63.3 -> 64.9Improved
tune k and p on PETS (ablation)ASIF · CIFAR100accuracy: 63.3 -> 60.9Hurt
tune k and p on CIFAR100 (ablation)ASIF · Petsaccuracy: 71.5 -> 63Hurt
sparsify relative representation to top k entries and exponentiate similarities by p (ablation)ASIF · zero-shot classificationaccuracyImproved
use smaller encoders in scaling experiment up to 1.6M pairsASIF · ImageNet zero-shotaccuracyNo clear effect

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy