In recent years, self-supervised learning has been studied to deal with the limitation of available labeled-dataset. Among the major components of self-supervised learning, the data augmentation pipeline is one key factor in enhancing the resulting performance. However, most researchers manually designed the augmentation pipeline, and the limited collections of transformation may cause the lack of robustness of the learned feature representation. In this work, we proposed Multi-Augmentations for Self-Supervised Representation Learning (MA-SSRL), which fully searched for various augmentation policies to build the entire pipeline to improve the robustness of the learned feature representation. MA-SSRL successfully learns the invariant feature representation and presents an efficient, effective, and adaptable data augmentation pipeline for self-supervised pre-training on different distribution and domain datasets. MA-SSRL outperforms the previous state-of-the-art methods on transfer and semi-supervised benchmarks while requiring fewer training epochs.
翻译:近年来,对自我监督学习进行了研究,以解决现有标签数据集的局限性问题,在自我监督学习的主要组成部分中,数据增强管道是提高由此而来的绩效的一个关键因素,然而,大多数研究人员手工设计了增强管道,而转型的收集有限,可能导致所学到的特征说明缺乏稳健性。在这项工作中,我们提出了“自我监督代表学习多重建议 ” ( MA-SSRL ), 充分寻找各种增强政策,以建设整个管道,提高学习特征代表的稳健性。 MA-SSRL 成功地学习了变化性特征代表,并展示了高效、有效和适应性的数据增强管道,用于对不同分布和域数据集进行自我监督的预先培训。 MA-SSRL 超越了先前最先进的传输方法和半超强基准,同时需要较少培训的用户。