Despite the remarkable performance of foundation vision-language models, the
shared representation space for text and vision can also encode harmful label
associations detrimental to fairness. While prior work has uncovered bias in
vision-language models' (VLMs) classification performance across geography,
work has been limited along the important axis of harmful label associations
due to a lack of rich, labeled data. In this work, we investigate harmful label
associations in the recently released Casual Conversations datasets containing
more than 70,000 videos. We study bias in the frequency of harmful label
associations across self-provided labels for age, gender, apparent skin tone,
and physical adornments across several leading VLMs. We find that VLMs are
$4-13$x more likely to harmfully classify individuals with darker skin tones.
We also find scaling transformer encoder model size leads to higher confidence
in harmful predictions. Finally, we find improvements on standard vision tasks
across VLMs does not address disparities in harmful label associations.

尽管基础视觉语言模型的性能出色，但文本和图像的共享表示空间也可能编码有害的标签关联，对公平性产生不利影响。本研究通过对最近发布的包含超过 70,000 个视频的 “Casual Conversations” 数据集进行调查，发现视觉语言模型在年龄、性别、肤色和装饰品等方面的有害标签关联频率存在偏差。我们发现，较暗肤色的个体更有可能被视觉语言模型误分类，并且发现随着变压器编码器模型规模的增加，对有害预测的置信度也增加。最后，我们发现在标准视觉任务上的改进并不能解决有害标签关联的不平等问题。

视觉语言模型中有害标签关联的偏见

The Bias of Harmful Label Associations in Vision-Language Models

How to learn highly compact yet effective sentence representation?
Pre-trained language models have been effective in many NLP tasks. However,
these models are often huge and produce large sentence embeddings. Moreover,
there is a big performance gap between large and small models. In this paper,
we propose Homomorphic Projective Distillation (HPD) to learn compressed
sentence embeddings. Our method augments a small Transformer encoder model with
learnable projection layers to produce compact representations while mimicking
a large pre-trained language model to retain the sentence representation
quality. We evaluate our method with different model sizes on both semantic
textual similarity (STS) and semantic retrieval (SR) tasks. Experiments show
that our method achieves 2.7-4.5 points performance gain on STS tasks compared
with previous best representations of the same size. In SR tasks, our method
improves retrieval speed (8.2$\times$) and memory usage (8.0$\times$) compared
with state-of-the-art large models.

这篇研究提出了 Homomorphic Projective Distillation（HPD）方法来学习压缩的句子嵌入，通过增加可学习的投影层来模拟大型预训练语言模型并保留句子表示质量。作者通过不同大小的模型对语义文本相似性（STS）和语义检索（SR）任务进行了评估，实验结果表明，相比于相同大小的最佳表示方法，该方法在 STS 任务上实现了 2.7-4.5 个性能提升，在 SR 任务上，该方法相比于最先进的大型模型，提高了检索速度（8.2 倍）和内存使用（8.0 倍）。