翻译：
What is the way AI (artificial intelligence) travels?
人工智能系统的运行方式是指AI系统的算力与存储资源所处的部署形态，构成上包括云端运行、设备端本地运行、边缘设备运行和混合云环境运行四种，解决的是不同场景下算力获取位置、响应时延与数据存放方式的选择问题，并不解决模型本身的算法精度与推理效果问题。
1. Run in the cloud: Running in the cloud is a common way of running AI (artificial intelligence). In this scenario, the computing and storage resources of the AI (artificial intelligence) system are located on cloud servers, and clients and other terminals are connected to the cloud via the Internet to access computing and storage resources. This model eliminates hardware and software facilities with high customer demand and ensures secure storage of data.
2. Run on the device side: Run on the device side is another mode of AI (artificial intelligence) system operation, also known as local operation. In this model, AI (artificial intelligence) systems run on devices (such as smartphones, smart speakers, etc.) rather than via cloud services. This pattern can make AI (artificial intelligence) systems more adaptable and support real-time response and interaction.
3. Run on edge devices: Run on edge devices refers to the execution of AI (artificial intelligence) systems at the position of the network close to the end device. It provides faster response times than cloud or on-premises operations and can support real-time data flow and analytics. Edge devices can be iot devices such as sensors, routers, base stations and smart homes.
4. Run in a hybrid cloud environment: A hybrid cloud environment refers to the integration of cloud computing and on-premises IT resources in a unified architecture in which AI (artificial intelligence) systems can run. In this model, AI (artificial intelligence) systems can perform distributed processing across multiple clouds and on-premises computers, achieving high availability, high performance, better security, and better maintenance.
In general, AI (artificial intelligence) systems can run in cloud, on-premises, edge devices, or hybrid cloud environments. Each mode of operation has its own advantages and disadvantages, and is suitable for different application scenarios and tasks. As users and developers of AI (artificial intelligence) technology, it is necessary to understand the characteristics and scope of different operating modes in order to make more informed choices for business decisions.
In artificial intelligence (AI) applications, data is a very important resource, and it is the basis for building machine learning models. Below is an overview of the types and sources of data.
