arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2201.10972cs.CVcs.AI
收藏

判别式训练的零样本学习模型的鲁棒性如何?

How Robust are Discriminatively Trained Zero-Shot Learning Models?

Hacettepe University(哈斯特佩大学);Middle East Technical University(中东技术大学)

Mehmet Kerim Yucel, Ramazan Gokberk Cinbis, Pinar Duygulu

摘要:

数据移动鲁棒性主要是从完全监督的角度进行研究的,而零样本学习(ZRL)模型的鲁棒性在很大程度上被忽略了。在本文中,我们提出了新的分析的鲁棒性的歧视性反卷积图像损坏。我们将几个CNOL模型置于一个大的常见腐败和防御集合中。为了实现腐败分析,我们策划并发布了第一个CRTL腐败鲁棒性数据集SUN-C,CUB-C和AWA 2-C。我们分析我们的结果,考虑到数据集的特点,类的不平衡,可见和不可见的类之间的类的转换和之间的差异,GALML和GALML性能。我们的研究结果表明,歧视性的CNOL遭受腐败,这种趋势进一步加剧了严重的类不平衡和CNOL方法固有的模型弱点。然后,我们将我们的研究结果与基于对抗攻击的结果结合起来,并强调了腐败和对抗示例的不同影响,例如对抗攻击下存在的伪鲁棒性效应。我们还获得了新的强大的基线与防御方法的两个模型。最后,我们的实验表明,虽然现有的方法,以提高鲁棒性的一些工作的CNOL模型,他们没有产生切实的效果。

英文摘要:

Data shift robustness has been primarily investigated from a fully supervised perspective, and robustness of zero-shot learning (ZSL) models have been largely neglected. In this paper, we present novel analyses on the robustness of discriminative ZSL to image corruptions. We subject several ZSL models to a large set of common corruptions and defenses. In order to realize the corruption analysis, we curate and release the first ZSL corruption robustness datasets SUN-C, CUB-C and AWA2-C. We analyse our results by taking into account the dataset characteristics, class imbalance, class transitions between seen and unseen classes and the discrepancies between ZSL and GZSL performances. Our results show that discriminative ZSL suffers from corruptions and this trend is further exacerbated by the severe class imbalance and model weakness inherent in ZSL methods. We then combine our findings with those based on adversarial attacks in ZSL, and highlight the different effects of corruptions and adversarial examples, such as the pseudo-robustness effect present under adversarial attacks. We also obtain new strong baselines for both models with the defense methods. Finally, our experiments show that although existing methods to improve robustness somewhat work for ZSL models, they do not produce a tangible effect.

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。

↑