Enhancing personalized modeling via weighted and adversarial learning
Enhancing personalized modeling via weighted and adversarial learning
复制标题
DOI：
10.1007/s41060-021-00263-3
复制
发表时间：
2021-05
影响因子：
2.4
通讯作者：
Wei Du;Xintao Wu
中科院分区：
文献类型：
--
作者：
Wei Du;Xintao Wu
在过去的几年中，数据产生源不断增加，例如移动的设备、嵌入式传感器、各种智能设备等。这些不断增加的数据源推动了深度学习模型的分布式部署。然而，传统的分布式深度学习是在所有收集的数据上建立一个全局模型，可能会忽略对个人用户至关重要的特定组件。在本文中，我们提出了一个对抗性学习框架，允许个人用户构建个性化模型。我们的框架由两个阶段组成，包括从其他用户中有效选择相似数据和对抗训练。我们不是通过计算手工设计的相似性度量来选择相似的数据，而是在单个用户的数据上训练自动编码器和生成对抗网络（GAN），并使用它们来请求其他用户的相似数据。为了进一步提高个性化模型的性能，我们开发了两种方法，结合联合收割机请求的数据和用户自己的数据来构建个性化模型。第一种方法是我们应用加权学习来捕获所请求数据的不同重要性。第二种方法是我们应用对抗训练来最小化请求数据和用户自己数据之间的分布差异。实验结果证明了该框架的有效性。
The data generation sources are increasing in the past few years, such as mobile devices, embedded sensors, various intelligent equipment and so forth. These increasing data sources push the deployment of deep learning models in a distributed manner. However, the traditional distributed deep learning is to build a global model over all collected data and may overlook specific components which are of vital importance to individual users. In this paper, we propose an adversarial learning framework that allows an individual user to build a personalized model. Our framework consists of two stages, including efficient similar data selection from other users and adversarial training. Instead of selecting similar data by computing hand-designed similarity metrics, we train an auto-encoder and a generative adversarial network (GAN) on individual user’s data and use them to request similar data from other users. To further improve the personalized model performance, we develop two approaches that combine the requested data and user’s own data to build the personalized model. The first approach is that we apply weighted learning to capture the different importance of the requested data. The second approach is that we apply adversarial training to minimize the distribution discrepancy between the requested data and user’s own data. Experimental results demonstrate the effectiveness of the proposed framework.