Vision-language models, such as CLIP, have shown promising Out-of-Distribution (OoD) generalization under various types of distribution shifts. Recent studies attempted to investigate the leading cause of this capability. In this work, we follow the same path, but focus on a specific type of OoD data - images with novel compositions of attribute-object pairs - and study whether such models can successfully classify those images into composition classes. We carefully designed an authentic image test dataset called ImageNet-AO, consisting of attributes for objects that are unlikely encountered in the CLIP training sets. We found that CLIPs trained with large datasets such as OpenAI CLIP, LAION-400M, and LAION-2B show orders-of-magnitude improvement in effective compositional OoD generalization compared to both supervised models and CLIPs trained with smaller datasets, such as CC-12M and YFCC-15M. Our results provide evidence that the scale and diversity of training data and language supervision play a key role in unlocking the compositional generalization abilities of vision-language models.
Duke University（杜克大学）；Tsinghua University（清华大学）；Microsoft Research（微软研究院）
A Semantic Compositional Network (SCN) is developed for image captioning, in which semantic concepts (i.e., tags) are detected from the image, and the probability of each tag is used to compose the parameters in a long short-term memory (LSTM) network. The SCN extends each weight matrix of the LSTM to an ensemble of tag-dependent weight matrices. The degree to which each member of the ensemble is used to generate an image caption is tied to the image-dependent probability of the corresponding tag. In addition to captioning images, we also extend the SCN to generate captions for video clips. We…
Michigan State University（密歇根州立大学）；University of Michigan（密歇根大学）
This work explores the zero-shot compositional learning ability of large pre-trained vision-language models(VLMs) within the prompt-based learning framework and propose a model (\textit{PromptCompVL}) to solve the compositonal zero-shot learning (CZSL) problem. \textit{PromptCompVL} makes two design choices: first, it uses a soft-prompting instead of hard-prompting to inject learnable parameters to reprogram VLMs for compositional learning. Second, to address the compositional challenge, it uses the soft-embedding layer to learn primitive concepts in different combinations. By combining both s…
Graph-guided Cross-composition Feature Disentanglement for Compositional Zero-shot Learning
PowerChina Huadong Engineering Corporation Limited（国家电力投资集团华东工程公司）；Hangzhou Dianzi University（杭州电子科技大学）；The University of Manchester（曼彻斯特大学）；The Hong Kong University of Science and Technology (Guangzhou)（香港科技大学（广州））；Nanjing University of Aeronautics and Astronautics（南京航空航天大学）；Zhejiang University（浙江大学）
Cross-Modal Masked Compositional Concept Modeling for Enhancing Visio-Linguistic Compositionality
MoE Key Laboratory of Brain-inspired Intelligent Perception and Cognition, University of Science and Technology of China（中国科学技术大学，教育部脑启发智能感知与认知重点实验室）；Independent Researcher（独立研究员）
Does Data Scaling Lead to Visual Compositional Generalization?
Tübingen AI Center, University of Tübingen（图宾根人工智能中心，图宾根大学）；Technical University of Munich（慕尼黑技术大学）；Munich Center for Machine Learning (MCML)（慕尼黑机器学习中心）；Max Planck Institute for Intelligent Systems, Tübingen（马克斯·普朗克智能系统研究所，图宾根）
While recent deep neural network models have achieved promising results on the image captioning task, they rely largely on the availability of corpora with paired image and sentence captions to describe objects in context. In this work, we propose the Deep Compositional Captioner (DCC) to address the task of generating descriptions of novel objects which are not present in paired image-sentence datasets. Our method achieves this by leveraging large object recognition datasets and external text corpora and by transferring knowledge between semantically similar concepts. Current deep caption mod…
Google Research（谷歌研究院）；Tel Aviv University（特拉维夫大学）；Weizmann Institute of Science（魏茨曼科学研究所）
Large vision-language models (VLMs), such as CLIP, learn rich joint image-text representations, facilitating advances in numerous downstream tasks, including zero-shot classification and text-to-image generation. Nevertheless, existing VLMs exhibit a prominent well-documented limitation - they fail to encapsulate compositional concepts such as counting. We introduce a simple yet effective method to improve the quantitative understanding of VLMs, while maintaining their overall performance on common benchmarks. Specifically, we propose a new counting-contrastive loss used to finetune a pre-trai…
Fairness Analysis of CLIP-Based Foundation Models for X-Ray Image Classification
Clinical Medical Research Center, The First People’s Hospital of Kashi (Kashgar) Prefecture（喀什地区第一人民医院临床医学研究中心）；Xinjiang Key Laboratory of Artificial Intelligence Assisted Imaging Diagnosis（新疆人工智能辅助影像诊断重点实验室）；Department of Hepatobiliary Surgery, The First People’s Hospital of Kashi (Kashgar) Prefecture（喀什地区第一人民医院肝胆外科）；Shanghai Artificial Intelligence Laboratory（上海人工智能实验室）
CLIP-Guided Adaptable Self-Supervised Learning for Human-Centric Visual Tasks
Key Laboratory of Intelligent Information Processing, Institute of Computing Technology (ICT), Chinese Academy of Sciences (CAS), and University of Chinese Academy of Sciences（智能信息处理重点实验室，计算技术研究所（ICT），中国科学院（CAS）及中国科学院大学）
Visual anomaly classification and segmentation are vital for automating industrial quality inspection. The focus of prior research in the field has been on training custom models for each quality inspection task, which requires task-specific images and annotation. In this paper we move away from this regime, addressing zero-shot and few-normal-shot anomaly classification and segmentation. Recently CLIP, a vision-language model, has shown revolutionary generality with competitive zero-/few-shot performance in comparison to full-supervision. But CLIP falls short on anomaly classification and seg…
