To answer a question from a classmate: In the current era of deep learning boom, is the expert system still useful? Will it be eliminated?

Insert image description here


My opinion is as follows: (It will not be eliminated, it will gradually evolve)

  1. 专家系统和深度学习有其各自的优势. Expert systems use rules and knowledge bases to give conclusions, and are suitable for areas with clear problem scope and clear rules. Deep learning learns patterns through large amounts of data training, and is suitable for areas where the amount of data is large but the rules are difficult to clarify.

  2. 两者不会互相取代,会结合使用. For example, deep learning is first used for preliminary screening, and then handed over to an expert system for rule verification and interpretation to achieve human-machine collaboration.

  3. 在部分领域,专家系统仍然很重要, such as medical diagnosis, legal judgment and other scenarios that require clear explanations of reasons. Deep learning black box models cannot be replaced.

  4. 专家系统会转型为知识图谱, integrated with deep learning and other technologies to provide more powerful decision support.


Summarize

So generally speaking, the expert system does not会被深度学习完全取代. But it needs to continuously absorb new technologies to improve functions and clarify their respective application scenarios. Both will continue to develop in different fields and will be the trend in the future. Expert systems are still in great demand. 人机协作解释能力的场景中发挥重要作用

Guess you like

Origin blog.csdn.net/qlkaicx/article/details/135005224