arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2304.03431cs.LGcs.AI
收藏

鲁棒不变表示中的域泛化

Domain Generalization In Robust Invariant Representation

Massachusetts Institute of Technology(麻省理工学院);University of California San Diego(加州大学圣迭戈分校);Thapar Institute of Engineering and Technology(塔帕尔工程技术学院)

Gauri Gupta, Ritvik Kapila, Keshav Gupta, Ramesh Raskar

AI总结:

本文研究不变表示在分布外数据上的泛化能力,通过实验证明对已见域变换保持不变的模型能学到对分布偏移鲁棒的非结构化潜在表示,使不变性成为资源受限训练中的理想性质。

摘要:

用于学习对常见变换不变表示的无监督方法在目标识别中被广泛使用。学习不变性使模型更加鲁棒,并在现实场景中更实用。由于不改变目标内在属性的数据变换造成了识别任务中的大部分复杂性,因此对这些变换具有不变性的模型有助于减少所需的训练数据量。这进一步提高了模型的效率并简化了训练。在本文中,我们研究了不变表示在分布外数据上的泛化能力,并试图回答以下问题:在某个特定已见域中对某些变换保持不变的模型表示,在先前未见过的域中是否也仍然保持不变?通过大量实验,我们证明该不变模型学习到的非结构化潜在表示对分布偏移具有鲁棒性,从而使不变性成为在资源受限环境下进行训练的一种理想性质。

英文摘要:

Unsupervised approaches for learning representations invariant to common transformations are used quite often for object recognition. Learning invariances makes models more robust and practical to use in real-world scenarios. Since data transformations that do not change the intrinsic properties of the object cause the majority of the complexity in recognition tasks, models that are invariant to these transformations help reduce the amount of training data required. This further increases the model's efficiency and simplifies training. In this paper, we investigate the generalization of invariant representations on out-of-distribution data and try to answer the question: Do model representations invariant to some transformations in a particular seen domain also remain invariant in previously unseen domains? Through extensive experiments, we demonstrate that the invariant model learns unstructured latent representations that are robust to distribution shifts, thus making invariance a desirable property for training in resource-constrained settings.

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。

↑