paper-with-me

Papers

TripletCLIP: Improving Compositional Reasoning of CLIP via Synthetic Vision-Language Negatives

2024-11-04 · Maitreya Patel, Abhiram Kusumba, Sheng Cheng, Changhoon Kim, Tejas Gokhale, Chitta Baral, Yezhou Yang

Contrastive Language-Image Pretraining (CLIP) models maximize the mutual information between text and visual modalities to learn representations. This makes the nature of the training data a significant factor in the efficacy of CLIP for downstream tasks. However, the lack of compositional diversity in contemporary image-text datasets limits the compositional reasoning ability of CLIP. We show that generating ``hard'' negative captions via in-context learning and synthesizing corresponding negative images with text-to-image generators offers a solution. We introduce a novel contrastive pre-training strategy that leverages these hard negative captions and images in an alternating fashion to train CLIP. We demonstrate that our method, named TripletCLIP, when applied to existing datasets such as CC3M and CC12M, enhances the compositional capabilities of CLIP, resulting in an absolute improvement of over 9% on the SugarCrepe benchmark on an equal computational budget, as well as improvements in zero-shot image classification and image retrieval. Our code, models, and data are available at: https://tripletclip.github.io

📄 PDF Abstract BibTeX arXiv:2411.02545

Code (0)

등록된 구현이 없습니다.

Tasks

Diversityimage-classificationImage ClassificationImage RetrievalIn-Context LearningZero-Shot Image Classification

Methods 이 논문이 사용한 방법론

CLIP Contrastive Language-Image Pre-training (CLIP), consisting of a simplified version of ConVIRT trained from scratch, is an efficient method of image representation learning…

Similar Papers 제목 키워드 기반

Does CLIP Bind Concepts? Probing Compositionality in Large Image Models

2022-12-20 · Martha Lewis, Nihal V. Nayak, Peilin Yu, Qinan Yu 외

Large-scale neural network models combining text and images have made incredible progress in recent years. However, it remains an open question to what extent such models encode compositional representations of the conce…

Language ModelingLanguage ModellingObjectOpen-Ended Question Answering

ReasonCLIP-58M: Visually Grounded Commonsense Reasoning Supervision for CLIP

2026-06-25 · Sicheng Zhang, Muzammal Naseer, Binzhu Xie, Naufal Suryanto 외 arxiv

CLIP and its variants are widely adopted visual backbones in multimodal systems, but their pretraining remains dominated by descriptive image-text alignment. As downstream applications increasingly demand visually ground…

Continual Pretraining

MLLMCLIP: Feature-Level Distillation of MLLM for Robust Vision-Language Representations

2026-08-26 · Jongsuk Kim, Qiyu Wu, Zhuoyuan Mao, Hiromi Wakaki 외 arxiv

Pretrained vision-language models such as CLIP excel at zero-shot recognition but often fail at compositionality, particularly attribute-object and relational structures. Recent studies mitigate this issue by augmenting …

Text Retrieval

Advancing Compositional Awareness in CLIP with Efficient Fine-Tuning

2025-05-30 · Amit Peleg, Naman Deep Singh, Matthias Hein

Vision-language models like CLIP have demonstrated remarkable zero-shot capabilities in classification and retrieval. However, these models often struggle with compositional reasoning - the ability to understand the rela…

Retrieval

Preserving Multi-Modal Capabilities of Pre-trained VLMs for Improving Vision-Linguistic Compositionality

2024-10-07 · Youngtaek Oh, Jae Won Cho, Dong-Jin Kim, In So Kweon 외

In this paper, we propose a new method to enhance compositional understanding in pre-trained vision and language models (VLMs) without sacrificing performance in zero-shot multi-modal tasks. Traditional fine-tuning appro…