A Flexible Selection Scheme for Minimum-Effort Transfer Learning
A Flexible Selection Scheme for Minimum-Effort Transfer Learning
复制标题
DOI：
10.1109/wacv45572.2020.9093635
复制
发表时间：
2020-03
期刊：
影响因子：
--
通讯作者：
Amélie Royer;Christoph H. Lampert
中科院分区：
文献类型：
--
作者：
Amélie Royer;Christoph H. Lampert
微调是一种利用预训练卷积网络中包含的知识进行新视觉识别任务的流行方式。然而，很少考虑将知识从预先训练的网络转移到视觉上不同但语义上接近的源的正交设置：这通常发生在现实生活中的数据中，这些数据不一定像训练源那样干净（噪声，几何变换，不同的模态等）。为了解决这种情况，我们引入了一种新的通用形式的微调，称为灵活调整，其中网络的任何单个单元（例如层）都可以进行调整，并且自动选择最有希望的单元。为了使该方法吸引实际使用，我们提出了两个轻量级和更快的选择程序，证明是很好的近似在实践中。我们研究这些选择标准的经验，在各种领域的转变和数据稀缺的情况下，并表明，微调个别单位，尽管其简单性，产生非常好的结果作为一种适应技术。事实证明，与通常的做法相反，在许多域转移场景中，最好调整中间或早期单元，而不是最后一个完全连接的单元，这可以通过灵活调整来准确检测。
Fine-tuning is a popular way of exploiting knowledge contained in a pre-trained convolutional network for a new visual recognition task. However, the orthogonal setting of transferring knowledge from a pretrained network to a visually different yet semantically close source is rarely considered: This commonly happens with real-life data, which is not necessarily as clean as the training source (noise, geometric transformations, different modalities, etc.).To tackle such scenarios, we introduce a new, generalized form of fine-tuning, called flex-tuning, in which any individual unit (e.g. layer) of a network can be tuned, and the most promising one is chosen automatically. In order to make the method appealing for practical use, we propose two lightweight and faster selection procedures that prove to be good approximations in practice. We study these selection criteria empirically across a variety of domain shifts and data scarcity scenarios, and show that fine-tuning individual units, despite its simplicity, yields very good results as an adaptation technique. As it turns out, in contrast to common practice, rather than the last fully-connected unit it is best to tune an intermediate or early one in many domain- shift scenarios, which is accurately detected by flex-tuning.