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 2201.01046cs.CVcs.MM
收藏

多预训练任务下的声音和视觉表示学习

Sound and Visual Representation Learning with Multiple Pretraining Tasks

ETH Zurich(苏黎世联邦理工学院);MPI for Informatics(马克斯·普朗克信息学研究所);KU Leuven(鲁汶大学)

Arun Balajee Vasudevan, Dengxin Dai, Luc Van Gool

摘要:

不同的自我监督任务(SSL)揭示了不同的数据特征。所学习的特征表示可以针对每个下游任务表现出不同的性能。在这种情况下,这项工作的目的是结合多个SSL任务(多SSL),以及一般化的所有下游任务。具体来说,在这项研究中,我们调查双耳声音和图像数据隔离。对于双耳声音,我们提出了三个SSL任务,即空间对齐,前景对象的时间同步和双耳音频和时间间隙预测。我们研究了多SSL的几种方法,并深入了解OmniAudio数据集上的视频检索,空间声音超分辨率和语义预测的下游任务性能。我们对双耳声音表示的实验表明,通过SSL任务的增量学习(IL)的多SSL优于单SSL任务模型和完全监督模型的下游任务性能。作为对其他模态的适用性检查,我们还制定了我们的多SSL模型的图像表示学习,我们使用最近提出的SSL任务,MoCov 2和DenseCL。在这里,Multi-SSL在VOC 07分类上超过MoCov 2,DenseCL和DetCo等最新方法2.06%,3.27%和1.19%,在COCO检测上超过+2.83,+1.56和+1.61 AP。代码将公开提供。

英文摘要:

Different self-supervised tasks (SSL) reveal different features from the data. The learned feature representations can exhibit different performance for each downstream task. In this light, this work aims to combine Multiple SSL tasks (Multi-SSL) that generalizes well for all downstream tasks. Specifically, for this study, we investigate binaural sounds and image data in isolation. For binaural sounds, we propose three SSL tasks namely, spatial alignment, temporal synchronization of foreground objects and binaural audio and temporal gap prediction. We investigate several approaches of Multi-SSL and give insights into the downstream task performance on video retrieval, spatial sound super resolution, and semantic prediction on the OmniAudio dataset. Our experiments on binaural sound representations demonstrate that Multi-SSL via incremental learning (IL) of SSL tasks outperforms single SSL task models and fully supervised models in the downstream task performance. As a check of applicability on other modality, we also formulate our Multi-SSL models for image representation learning and we use the recently proposed SSL tasks, MoCov2 and DenseCL. Here, Multi-SSL surpasses recent methods such as MoCov2, DenseCL and DetCo by 2.06%, 3.27% and 1.19% on VOC07 classification and +2.83, +1.56 and +1.61 AP on COCO detection. Code will be made publicly available.

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。

↑