Open AccessUsing MLOPS framework to increase machine learning life cycle in IоT
Author(s) -
O. M. Nikolayenko,
AUTHOR_ID,
O. V. Polonevych
Publication year - 2021
Publication title -
zv'âzok
Language(s) - English
Resource type - Journals
ISSN - 2412-9070
DOI - 10.31673/2412-9070.2021.024952
Subject(s) - edge computing , cloud computing , computer science , edge device , enhanced data rates for gsm evolution , big data , software deployment , boom , process (computing) , distributed computing , container (type theory) , artificial intelligence , deep learning , data science , software engineering , data mining , engineering , operating system , mechanical engineering , environmental engineering
Recent years have witnessed a boom in IoT devices, leading to high data volumes and low latency demand, which has led to demand for the 5G networks. This infrastructure shift allows for real-time decision-making using artificial intelligence for IoT applications. Artificial intelligence (AIoT) is a combination of artificial intelligence (AI) technology with the Internet of Things infrastructure (IoT) to achieve more efficient IoT operations and decision making. Edge computing is emerging to enable AIoT applications. Edge computing enables generating insights and making decisions at the data source, reducing the amount of data sent to the cloud and central repository. An ecosystem to facilitate edge computing for AIoT applications has become essential to make real-time decisions. Edge computing is the process of performing computational tasks that are physically close to the target devices, rather than in the cloud or on the device itself. This allows you to gain knowledge, ideas, and decisions at the data source. The purpose of edge computing should bring the calculations closer to the data source and unload the centralized calculations to the decentralized ones. Edge computing allows the application of various machine learning algorithms to create new experiences and new opportunities in many industries, from a connected house to security, surveillance, and automotive. As for MLOps, IoT Edge is another deployment platform. However, if we use models on the IoT Edge, we need to consider some additional considerations. IoT Edge-targeted MLOps models must run offline mode. IoT models are more susceptible to data drift due to high data rates. IoT machine learning models need to be deployed on different target platforms, and we need to use the capabilities of these platforms.