By: Vajratiya Vajrobol, International Center for AI and Cyber Security Research and Innovations (CCRI), Asia University, Taiwan, vvajratiya@gmail.com
Two novel ideas that are redefining data processing and machine learning model training in the technology era are Edge Computing and Federated Learning. Both ideas present a decentralised computing strategy.In this article, we will dive into differences in two techniques.
- Edge Computing
Processing data near to the data source, sometimes referred to as the “edge” of the network, is the basis of the edge computing concept. Edge computing, in contrast to conventional centralised cloud computing, uses decentralised computing resources to improve real-time processing and lower latency [1] . It works especially well for applications that need to react quickly, including those that use sensors or Internet of Things (IoT) devices [2]. By processing data locally, edge computing reduces the quantity of data that needs to be transmitted across the network, improving bandwidth efficiency.
- Federated Learning:
Federated learning is a machine learning technique that focuses on decentralised model training across multiple devices or edge servers. Federated learning transmits and aggregates just model updates—not raw data—from each device where the model is trained locally [3] . The goal of this collaborative learning approach is to mitigate privacy concerns by sharing only aggregated model updates and retaining sensitive data locally. Federated learning enables collaborative model training without jeopardising the privacy of individual data sources, making it especially pertinent in settings where privacy is a top concern, such as healthcare, finance [4,5].
- Key Differences :
Although distributed computing is a component of both edge computing and federated learning, its main goals and traits are different. The main goal of edge computing is local data processing, which lowers latency and makes it appropriate for real-time applications. Federated learning, on the other hand, focuses on decentralised model training, sending just model updates and maintaining local data, thereby putting privacy first. While federated learning combines decentralised training with a centralised model aggregation step, edge computing frequently involves decentralised processing. The two ideas can work in situations where cooperative model training and local processing are both critical, providing a comprehensive approach to distributed computing across a range of industries [6].
In summary, the partnership between Federated Learning and Edge Computing offers a bright future for distributed computing. Federated Learning, which prioritises privacy by utilising decentralised devices to train machine learning models collectively, is in perfect harmony with Edge Computing’s focus on local data processing for lower latency.
References
- Cao, K., Liu, Y., Meng, G., & Sun, Q. (2020). An overview on edge computing research. IEEE access, 8, 85714-85728.
- Li, H., Ota, K., & Dong, M. (2018). Learning IoT in edge: Deep learning for the Internet of Things with edge computing. IEEE network, 32(1), 96-101.
- Li, L., Fan, Y., Tse, M., & Lin, K. Y. (2020). A review of applications in federated learning. Computers & Industrial Engineering, 149, 106854.
- Long, G., Tan, Y., Jiang, J., & Zhang, C. (2020). Federated learning for open banking. In Federated Learning: Privacy and Incentive (pp. 240-254). Cham: Springer International Publishing.
- Antunes, R. S., André da Costa, C., Küderle, A., Yari, I. A., & Eskofier, B. (2022). Federated learning for healthcare: Systematic review and architecture proposal. ACM Transactions on Intelligent Systems and Technology (TIST), 13(4), 1-23.
- Abreha, H. G., Hayajneh, M., & Serhani, M. A. (2022). Federated learning in edge computing: a systematic survey. Sensors, 22(2), 450.
Cite As:
Vajrobol V. (2024) The differences between Edge Computing and Federated Learning, Insights2Techinfo, pp.1