Skip to yearly menu bar Skip to main content


Poster

Improving fine-grained understanding in image-text pre-training

Ioana Bica · Anastasija Ilic · Matthias Bauer · Goker Erdogan · Matko BoĆĄnjak · Christos Kaplanis · Alexey Gritsenko · Matthias Minderer · Charles Blundell · Razvan Pascanu · Jovana Mitrovic

Hall C 4-9 #708
[ ] [ Paper PDF ]
[ Poster
Thu 25 Jul 2:30 a.m. PDT — 4 a.m. PDT

Abstract:

We introduce SPARse fine-grained Contrastive alignment (SPARC), a simple method for pretraining more fine-grained multimodal representations from image-text pairs. Given that multiple image patches often correspond to single words, we propose to learn a grouping of image patches for every token in the caption. To achieve this, we use a sparse similarity metric between image patches and language tokens and compute for each token a language-grouped vision embedding as the weighted average of patches. The token and language-grouped vision embeddings are then contrasted through a fine-grained sequence-wise loss that only depends on individual samples and does not require other batch samples as negatives, i.e., more detailed information is encoded in a computationally inexpensive way. SPARC combines this fine-grained loss with a contrastive loss between global image and text embeddings to learn representations that simultaneously encode global and local information. We thoroughly evaluate SPARC and show improved performance over competing approaches both on image-level tasks relying on coarse-grained information, e.g. classification, as well as region-level tasks relying on fine-grained information, e.g., retrieval, object detection, segmentation while also improving model faithfulness and captioning in foundational vision-language models.

Chat is not available.