Local Aggregation for Unsupervised Learning of Visual Embeddings
Chengxu Zhuang Stanford University Alex Lin Zhai Stanford University Daniel Yamins Stanford University
Abstract
Unsupervised approaches to learning in neural networks are of substantial interest for furthering artificial intelligence, both because they would enable the training of networks without the need for large numbers of expensive annotations, and because they would be better models of the kind of general-purpose learning deployed by humans. However, unsupervised networks have long lagged behind the performance of their supervised counterparts, especially in the domain of large-scale visual recognition. Recent developments in training deep convolutional embeddings to maximize non-parametric instance separation and clustering objectives have shown promise in closing this gap. Here, we describe a method that trains an embedding function to maximize a metric of local aggregation, causing similar data instances to move together in the embedding space, while allowing dissimilar instances to separate. This aggregation metric is dynamic, allowing soft clusters of different scales to emerge. We evaluate our procedure on several large-scale visual recognition datasets, achieving state-of-the-art unsupervised transfer learning performance on object recognition in ImageNet, scene recognition in P
中文速览
大规模视觉识别任务长期依赖大量人工标注数据，无监督学习虽能省去标注，但效果与监督学习相比差距明显。研究者提出了一种叫做"局部聚合"（Local Aggregation，LA）的无监督训练方法：让神经网络把相似的图像在嵌入空间（embedding space）中自动靠拢、把不相似的推开，通过动态识别每张图像的"近邻"和"背景邻居"来度量聚类程度，并以此作为优化目标迭代训练，不需要任何类别标签。在ImageNet、Places 205和PASCAL VOC等主流大规模视觉数据集上的实验表明，LA在无监督迁移学习（transfer learning）性能上超越了此前所有方法，ImageNet分类Top-1准确率达到60.2%，首次超过了直接用监督方式训练的AlexNet基准。这项工作证明，无监督方法已经可以在实际大规模视觉任务中媲美甚至超过经典监督模型，为减少对昂贵标注数据的依赖提供了切实可行的新路径。
原文 arXiv:1903.12355;中英对照 + 大白话阅读 https://aha.fim.ai/paper/1903.12355v2