2019 AICON Seminar
This paper examines the effectiveness of applying deep learning through a variety of process (eg, medical and manufacturing steel mills, semiconductor processes, general chatbots, and financial industries) data.
The most important thing in deep learning is the legacy architecture of the system to be applied, and in the manufacturing industry, understanding the process is the most important.
This is because deep learning models applied to processes that improve efficiency and that can be directly linked to productivity gains can generate enormous cost savings. I don't know much about the steel industry I used to, but I compare the processes of the semiconductor industry and consider where deep learning algorithms might be applied.
In addition, in the medical industry, security is important above all, and Federated Learning should be used, and in the case of chatbots, an overall overview of the importance of architectural design according to the Intent Scope and the transfer learning will come out.
23. Basic process for producing water
Bituminous coal
Temperature
carbon monoxide (CO)
Time
impurities such as carbon and sulfu
Impurity concentration
Process of removing impurities from
water to make steel
This process filters out phosphorus,
sulfur and carbon.
Liquid iron becomes solid
It is injected into a mold, which is still
liquid, molten steel, cooled and solidified
through a continuous casting machine,
and continuously made into intermediate
materials such as slab, bloom and billet.
Process to be the source of final result
Rolling Process (์์ฐ๊ณต์ )
This rolling process is divided into two
types, hot rolling and cold rolling.
Continuous casting (์ฐ์ฃผ๊ณต์
)
Iron making (์ ์ ๊ณต์
)
Steel making (์ ๊ฐ๊ณต์ )