With the rapid development of large language models, with the popularity of ChatGPT, people slowly find what large models can do, but subject to some difficulties in central computing cloud computing, many people turn their attention to edge computing, the ideal state of edge computing is that people holding a mobile phone or an APP, can be used as an edge computing device, or in the face of intelligent humanoid robots, embodied robots for edge computing, and then interact with central computing, just like edge computing during the day, At night, central interaction, this is actually a bit like federated learning, and there is still a certain connection with federated learning, but the difference is that this is continuous mutual transmission, which is forward and reverse federated learning.
This method is also a very good way to connect large models and edge computing terminals in the future, in fact, this aspect is the preparation of hardware, and the other aspect is the preparation of edge computing algorithms.
Edge computing is a distributed computing model that aims to transfer computing resources and data processing functions from centralized cloud computing centers to edge devices close to the data source to improve the efficiency and real-time performance of data processing. In edge computing, data can be processed and analyzed on edge devices such as devices, sensors, and smart terminals, rather than sending it all to the cloud for processing. This reduces data transfer latency and network congestion, and improves data privacy and security. With the continuous development of large models and edge computing, a new type of intelligent application mode may emerge in the future, that is, the computing and data processing tasks of large models are assigned to intelligent terminals or edge servers for processing through edge computing to achieve more efficient and real-time intelligent applications. This intelligent application model will have higher autonomy and distributed characteristics, and can better adapt to various complex application scenarios.
In terms of hardware, in order to achieve the efficiency and real-time performance of edge computing, it is necessary to develop more advanced hardware devices such as chips and sensors. These hardware devices need to have high performance, low power consumption, low cost, and stable operation in various complex environments. In addition, more efficient algorithms and data processing technologies need to be investigated to achieve faster data processing and analysis tasks on edge devices.
In terms of algorithms, edge computing needs to research and apply more intelligent algorithms and technologies, such as artificial intelligence, machine learning, deep learning, etc. These algorithms and technologies can help edge devices better process and analyze data for more efficient and intelligent applications.
With the continuous development of large language models and edge computing technology, future intelligent applications will pay more attention to the characteristics of efficiency, autonomy and real-time. Through continuous research and application of new hardware and algorithm technologies, we can better promote the development and application of edge computing technology and bring broader development prospects for future intelligent applications. #内算与人工智能#