★Machine Learning Trends in 2022
In my 2nd article, I would like to share some of the ML trends that are introduced in the article I recently read, "Machine Learning Technology Trends To Impact Business in 2022"! (https://mobidev.biz/blog/future-machine-learning-trends-impact-business)
This time, I picked 2 trends below:
⑴ Edge AI ・TinyML
In recent years, the IoT (Internet of things) has become an important part of our daily lives. This has led to a focus on Edge AI (a type of Edge computing), which acquires and processes data on Edge devices without passing through the cloud as much as possible.
In the past, data acquired on Edge devices was sent to the cloud or on-premise servers without optimization, but this caused problems such as slow processing when the data was large.
Edge computing technology has made it possible to process applications in real-time, reduce the cost of data communication and processing in the cloud, and solve security risks associated with sending unprocessed data, leading to the widespread use of Edge AI products (e.g. Edge AI camera). Edge AI technology is expected to become even more indispensable in our daily lives in 2022.
⑵ MLOps
In recent years, the idea of DevOps has become quite popular in software development, which is simply the practice of automating the flow of operations and infrastructure to maintain the quality of the service/product and make it more scalable. MLOps is the application of this DevOps concept to the ML lifecycle (see below) to improve reliability and enable larger data to be handled.
___________
1, Design ML models to solve business problems
2, Acquisition, processing, and preparation of data for ML models
3, Train ML models
4, Validate the ML model
5, Actual deployment of the model
6, Monitoring the process to improve the ML model
___________
【Keyword】
I would like to also introduce K8s, a tool commonly used in both Edge AI and MLOps technologies.
・Kubernetes (K8s) : Kubernetes (K8s): K8s is an open source system (OSS) that helps manage and scale containers, often used with an application container called Docker to help containerize and automate processes.
Simply put, Docker is the container and K8s is the crane that deploys the container.
Containerisation is often used in DevOps practices and is a technology that allows applications to be more easily managed and deployed by packaging them in an environment called a 'container'. In recent years, there has been a practice to incorporate this technology into Edge computing, making it easier to manage and operate even after larger deployments or an increase in the number of edge devices.
【References】
・「What is edge computing?」https://youtu.be/cEOUeItHDdo
・「DockerとKubernetesの違いって何?」https://youtu.be/3c_sN3BaSwU
・「Kubernetesをエッジで実行するために」
https://avinton.com/blog/2022/05/how-to-run-kubernetes-on-the-edge/
この記事が気に入ったらサポートをしてみませんか?