MOFI: Learning Image Representations from Noisy Entity Annotated Images
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2023
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866929280251854848 |
|---|---|
| author | Wu, Wentao Timofeev, Aleksei Chen, Chen Zhang, Bowen Duan, Kun Liu, Shuangning Zheng, Yantao Shlens, Jonathon Du, Xianzhi Gan, Zhe Yang, Yinfei |
| author_facet | Wu, Wentao Timofeev, Aleksei Chen, Chen Zhang, Bowen Duan, Kun Liu, Shuangning Zheng, Yantao Shlens, Jonathon Du, Xianzhi Gan, Zhe Yang, Yinfei |
| contents | We present MOFI, Manifold OF Images, a new vision foundation model designed to learn image representations from noisy entity annotated images. MOFI differs from previous work in two key aspects: (i) pre-training data, and (ii) training recipe. Regarding data, we introduce a new approach to automatically assign entity labels to images from noisy image-text pairs. Our approach involves employing a named entity recognition model to extract entities from the alt-text, and then using a CLIP model to select the correct entities as labels of the paired image. It's a simple, cost-effective method that can scale to handle billions of web-mined image-text pairs. Through this method, we have created Image-to-Entities (I2E), a new dataset with 1 billion images and 2 million distinct entities, covering rich visual concepts in the wild. Building upon the I2E dataset, we study different training recipes like supervised pre-training, contrastive pre-training, and multi-task learning. For contrastive pre-training, we treat entity names as free-form text, and further enrich them with entity descriptions. Experiments show that supervised pre-training with large-scale fine-grained entity labels is highly effective for image retrieval tasks, and multi-task training further improves the performance. The final MOFI model achieves 86.66% mAP on the challenging GPR1200 dataset, surpassing the previous state-of-the-art performance of 72.19% from OpenAI's CLIP model. Further experiments on zero-shot and linear probe image classification also show that MOFI outperforms a CLIP model trained on the original image-text data, demonstrating the effectiveness of the I2E dataset in learning strong image representations. We release our code and model weights at https://github.com/apple/ml-mofi. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2306_07952 |
| institution | arXiv |
| publishDate | 2023 |
| record_format | arxiv |
| spellingShingle | MOFI: Learning Image Representations from Noisy Entity Annotated Images Wu, Wentao Timofeev, Aleksei Chen, Chen Zhang, Bowen Duan, Kun Liu, Shuangning Zheng, Yantao Shlens, Jonathon Du, Xianzhi Gan, Zhe Yang, Yinfei Computer Vision and Pattern Recognition Computation and Language Machine Learning We present MOFI, Manifold OF Images, a new vision foundation model designed to learn image representations from noisy entity annotated images. MOFI differs from previous work in two key aspects: (i) pre-training data, and (ii) training recipe. Regarding data, we introduce a new approach to automatically assign entity labels to images from noisy image-text pairs. Our approach involves employing a named entity recognition model to extract entities from the alt-text, and then using a CLIP model to select the correct entities as labels of the paired image. It's a simple, cost-effective method that can scale to handle billions of web-mined image-text pairs. Through this method, we have created Image-to-Entities (I2E), a new dataset with 1 billion images and 2 million distinct entities, covering rich visual concepts in the wild. Building upon the I2E dataset, we study different training recipes like supervised pre-training, contrastive pre-training, and multi-task learning. For contrastive pre-training, we treat entity names as free-form text, and further enrich them with entity descriptions. Experiments show that supervised pre-training with large-scale fine-grained entity labels is highly effective for image retrieval tasks, and multi-task training further improves the performance. The final MOFI model achieves 86.66% mAP on the challenging GPR1200 dataset, surpassing the previous state-of-the-art performance of 72.19% from OpenAI's CLIP model. Further experiments on zero-shot and linear probe image classification also show that MOFI outperforms a CLIP model trained on the original image-text data, demonstrating the effectiveness of the I2E dataset in learning strong image representations. We release our code and model weights at https://github.com/apple/ml-mofi. |
| title | MOFI: Learning Image Representations from Noisy Entity Annotated Images |
| topic | Computer Vision and Pattern Recognition Computation and Language Machine Learning |
| url | https://arxiv.org/abs/2306.07952 |