偏见进，偏见出？在原始模型中发现无偏子网络
Bias In, Bias Out? Finding Unbiased Subnetworks in Vanilla Models
提出通过剪枝从标准训练模型中提取无偏子网络，无需额外数据或重训练，为视觉模型去偏提供高效的结构化方法，可复用于其他公平性任务。
原文依据 such subnetworks can be extracted via pruning and can operate without modification
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论深度学习中的算法偏见问题促使了各种去偏技术的发展，其中许多方法需要复杂的训练过程或数据集操作。然而，一个有趣的问题出现了：是否有可能从标准原始训练的模型中提取公平且与偏见无关的子网络，而不依赖额外数据（如无偏训练集）？在本工作中，我们引入了偏见不变子网络提取（BISE），这是一种学习策略，能够识别并分离常规训练模型中已存在的“无偏”子网络，而无需重新训练或微调原始参数。我们的方法表明，这样的子网络可以通过剪枝提取，并且无需修改即可运行，有效减少对偏见特征的依赖并保持稳健性能。我们的发现有助于通过参数移除对预训练神经网络进行结构自适应，从而实现高效的偏见缓解，而不是采用以数据为中心或涉及（重新）训练所有模型参数的昂贵策略。在常见基准上的大量实验表明，我们的方法在所得去偏模型的性能和计算效率方面具有优势。
英文原文摘要 1261 字符
The issue of algorithmic biases in deep learning has led to the development of various debiasing techniques, many of which perform complex training procedures or dataset manipulation. However, an intriguing question arises: is it possible to extract fair and bias-agnostic subnetworks from standard vanilla-trained models without relying on additional data, such as unbiased training set? In this work, we introduce Bias-Invariant Subnetwork Extraction (BISE), a learning strategy that identifies and isolates "bias-free" subnetworks that already exist within conventionally trained models, without retraining or finetuning the original parameters. Our approach demonstrates that such subnetworks can be extracted via pruning and can operate without modification, effectively relying less on biased features and maintaining robust performance. Our findings contribute towards efficient bias mitigation through structural adaptation of pre-trained neural networks via parameter removal, as opposed to costly strategies that are either data-centric or involve (re)training all model parameters. Extensive experiments on common benchmarks show the advantages of our approach in terms of the performance and computational efficiency of the resulting debiased model.