"Deep learning is a technology that has already withered." Google professes that deep learning is not new. "It's out of the question that general companies can't decide whether to implement it now. Why not use it? Because it is a technology that is beginning to die (fully used)."
Conventionally, unstructured data such as images and sounds was difficult to be automatically judged, but deep learning has made it possible to automatically judge with high accuracy, and it is beginning to be used in the field of business.
In the example of a retail store, deep learning analyzes unstructured data such as "face data" to show not only the attributes of customers but also where in the store they are and what kind of action they are taking. .. From that, you can infer the reason why you did not buy.
Linking information on attributes and lead behavior with "purchasing information" of POS data, a tendency that "a person who took such lead behavior has a 65% chance of buying such a product" will be compiled into a database. .. By analyzing it, concrete measures can be planned.
It has become possible to grasp the behavior of customers in the retail business and to automate defective product inspection in factories that rely on human visual inspection. Furthermore, it will be possible to verify hypotheses based on data and make more accurate decisions, leading to profit generation.
AI should be used to solve the management issues that are the bottleneck of the company. Then, unstructured data (images, sounds, etc.) collected by deep learning becomes a treasure trove by introducing AI.
This makes it possible to verify hypotheses based on fact data and to study and plan policies without relying solely on experience and perception.
