Decentralized Knowledge Acquisition for Mobile Internet Applications
- PDF / 1,535,940 Bytes
- 17 Pages / 439.642 x 666.49 pts Page_size
- 55 Downloads / 212 Views
Decentralized Knowledge Acquisition for Mobile Internet Applications Jing Jiang1
· Shaoxiong Ji2
· Guodong Long1
Received: 17 March 2019 / Revised: 11 September 2019 / Accepted: 23 December 2019 / © Springer Science+Business Media, LLC, part of Springer Nature 2020
Abstract Mobile internet applications on smart phones dominate large portions of daily life for many people. Conventional machine learning-based knowledge acquisition methods collect users’ data in a centralized server, then train an intelligent model, such as recommendation and prediction, using all the collected data. This knowledge acquisition method raises serious privacy concerns, and also violates the rules of the newly published General Data Protection Regulation. This paper proposes a new attention-augmented federated learning framework that can conduct decentralized knowledge acquisition for mobile Internet application scenarios, such as mobile keyboard suggestions. In particular, the attention mechanism aggregates the decentralized knowledge which has been acquired from each mobile using its own data locally. The centralized server aggregates knowledge without direct access to personal data. Experiments on three real-world datasets demonstrate that the proposed framework performs better than other baseline methods in terms of perplexity and communication cost. Keywords Federated learning · Mobile internet applications · Decentralized knowledge acquisition
This article belongs to the Topical Collection: Special Issue on Application-Driven Knowledge Acquisition Guest Editors: Xue Li, Sen Wang, and Bohan Li Guodong Long
[email protected] Jing Jiang [email protected] Shaoxiong Ji [email protected] 1
Central for Artificial Intelligence, School of Computer Science, Faculty of Engineering and Information Technology, University of Technology Sydney, Ultimo, Australia
2
Department of Computer Science, Aalto University, Espoo, Finland
World Wide Web
1 Introduction According to a report in 2017 [29], the majority of respondents spent five hours or more a day on their smartphones daily. The increasing popularity of smartphones and mobile wearable devices, such as smart watches has generated a massive amount of distributed data, such as text messages, click behavior, travel trajectories and health status. Effectively using these data will provide additional advantages for these application service providers. Intelligent functions, such as recommendation and personalized suggestions, can greatly improve the quality of services and provide a more satisfying experience with the application. A conventional intelligent approach usually collects big data into a centralized server, then applies a machine learning algorithm to train an intelligent model. The personalized intelligent model is also trained in the centralized sever to leverage data from many other users. This centralized knowledge acquisition method raises serious privacy concerns for end users who have no idea as to how the company will use their data. One of the best ways
Data Loading...
