arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2204.12941cs.LGcs.CV
收藏

无偏视觉表示的无监督学习

Unsupervised Learning of Unbiased Visual Representations

University of Turin(都灵大学);Télécom Paris(巴黎电信学院);IP Paris(巴黎理工学院)

Carlo Alberto Barbano, Enzo Tartaglione, Marco Grangetto

AI总结:

针对数据集偏差导致模型泛化差且无偏差标注与先验知识的问题,提出完全无监督去偏差框架,含偏差捕捉、伪标签获取、有监督去偏差三步,结合理论框架分析,实验在合成与真实数据集上性能超部分有监督方法。

摘要:

深度神经网络在存在数据集偏差的情况下,往往难以学习到鲁棒的表示,导致在无偏数据集上的泛化性能欠佳。这一局限性源于模型在训练过程中会过度依赖外围和混杂因素,而这些因素是在训练中无意习得的。现有解决该问题的方法通常需要对偏差属性进行显式监督,或依赖于关于偏差的先验知识。在本研究中,我们针对无偏差显式标注且对其性质无先验知识的挑战性场景展开研究。我们提出了一个完全无监督的去偏差框架,包含三个关键步骤:首先,利用模型学习有害偏差的固有倾向来获取一个捕捉偏差的模型;其次,通过伪标签过程获得偏差标签;最后,应用前沿的有监督去偏差技术实现无偏模型。此外,我们引入了一个评估模型偏差的理论框架,并详细分析了偏差如何影响神经网络训练。在合成和真实世界数据集上的实验结果证明了我们方法的有效性,在多种设置下展现出最先进的性能,有时甚至超过了完全有监督的去偏差方法。

英文摘要:

Deep neural networks often struggle to learn robust representations in the presence of dataset biases, leading to suboptimal generalization on unbiased datasets. This limitation arises because the models heavily depend on peripheral and confounding factors, inadvertently acquired during training. Existing approaches to address this problem typically involve explicit supervision of bias attributes or reliance on prior knowledge about the biases. In this study, we address the challenging scenario where no explicit annotations of bias are available, and there's no prior knowledge about its nature. We present a fully unsupervised debiasing framework with three key steps: firstly, leveraging the inherent tendency to learn malignant biases to acquire a bias-capturing model; next, employing a pseudo-labeling process to obtain bias labels; and finally, applying cutting-edge supervised debiasing techniques to achieve an unbiased model. Additionally, we introduce a theoretical framework for evaluating model biasedness and conduct a detailed analysis of how biases impact neural network training. Experimental results on both synthetic and real-world datasets demonstrate the effectiveness of our method, showcasing state-of-the-art performance in various settings, occasionally surpassing fully supervised debiasing approaches.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑