paper-with-me

Papers

Vision-Language Models Performing Zero-Shot Tasks Exhibit Gender-based Disparities

2023-01-26 · Melissa Hall, Laura Gustafson, Aaron Adcock, Ishan Misra, Candace Ross

We explore the extent to which zero-shot vision-language models exhibit gender bias for different vision tasks. Vision models traditionally required task-specific labels for representing concepts, as well as finetuning; zero-shot models like CLIP instead perform tasks with an open-vocabulary, meaning they do not need a fixed set of labels, by using text embeddings to represent concepts. With these capabilities in mind, we ask: Do vision-language models exhibit gender bias when performing zero-shot image classification, object detection and semantic segmentation? We evaluate different vision-language models with multiple datasets across a set of concepts and find (i) all models evaluated show distinct performance differences based on the perceived gender of the person co-occurring with a given concept in the image and that aggregating analyses over all concepts can mask these concerns; (ii) model calibration (i.e. the relationship between accuracy and confidence) also differs distinctly by perceived gender, even when evaluating on similar representations of concepts; and (iii) these observed disparities align with existing gender biases in word embeddings from language models. These findings suggest that, while language greatly expands the capability of vision tasks, it can also contribute to social biases in zero-shot vision settings. Furthermore, biases can further propagate when foundational models like CLIP are used by other models to enable zero-shot capabilities.

📄 PDF Abstract BibTeX arXiv:2301.11100

Code (0)

등록된 구현이 없습니다.

Tasks

image-classificationImage Classificationobject-detectionObject DetectionSemantic SegmentationWord EmbeddingsZero-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…
ALIGN In the ALIGN method, visual and language representations are jointly trained from noisy image alt-text data. The image and text encoders are learned via contrastive loss…

Similar Papers 제목 키워드 기반

Filtering, Distillation, and Hard Negatives for Vision-Language Pre-Training

2023-01-05 · CVPR 2023 1 · Filip Radenovic, Abhimanyu Dubey, Abhishek Kadian, Todor Mihaylov 외

Vision-language models trained with contrastive learning on large-scale noisy data are becoming increasingly popular for zero-shot recognition problems. In this paper we improve the following three aspects of the contras…

Contrastive LearningText SpottingZero-Shot Learning

On the Zero-shot Adversarial Robustness of Vision-Language Models: A Truly Zero-shot and Training-free Approach

2025-01-01 · CVPR 2025 1 · Baoshun Tong, Hanjiang Lai, Yan Pan, Jian Yin

Pre-trained Vision-Language Models (VLMs) like CLIP, have demonstrated strong zero-shot generalization capabilities. Despite their effectiveness on various downstream tasks, they remain vulnerable to adversarial samp…

Adversarial RobustnessZero-shot Generalization

Let's Do a Thought Experiment: Using Counterfactuals to Improve Moral Reasoning

2023-06-25 · Xiao Ma, Swaroop Mishra, Ahmad Beirami, Alex Beutel 외

Language models still struggle on moral reasoning, despite their impressive performance in many other tasks. In particular, the Moral Scenarios task in MMLU (Multi-task Language Understanding) is among the worst performi…

counterfactualMathMMLUMoral Scenarios+1

ImageBind: One Embedding Space To Bind Them All

2023-05-09 · CVPR 2023 1 · Rohit Girdhar, Alaaeldin El-Nouby, Zhuang Liu, Mannat Singh 외

We present ImageBind, an approach to learn a joint embedding across six different modalities - images, text, audio, depth, thermal, and IMU data. We show that all combinations of paired data are not necessary to train su…

AllCross-Modal RetrievalMultimodal Deep LearningRetrieval+10

Modeling Caption Diversity in Contrastive Vision-Language Pretraining

2024-04-30 · Samuel Lavoie, Polina Kirichenko, Mark Ibrahim, Mahmoud Assran 외

There are a thousand ways to caption an image. Contrastive Language Pretraining (CLIP) on the other hand, works by mapping an image and its caption to a single vector -- limiting how well CLIP-like models can represent t…

Diversityzero-shot-classificationZero-Shot Learning