Fortune 500 companies 20 years achievements, unicorns are made in 4 years.

Through the new changes of digital transformation or the 4th Industrial Revolution, the current unicorn company has created the value of 1.4 trillion won in the past 20 years by the 500th Fortune companies in the past. Companies are leading the industry's innovation through cloud computing.

Google Cloud Korea Park Hyung-jun, a customer engineer, said at the 19th Advanced Computing Conference+ 2022 held online.

Park Hyung-jun announced the theme of 'Digital Transformation with Google'.

He suggested how digital transformation changes in everyday life and industry, detecting safety threats at the construction site, automating visual inspection at the manufacturing site, and supporting decision-making of medical staff. Cases such as predicting the quantity of raw materials in the retail distribution industry, SCM efficiency, personalized services, and using the financial service industry's funds, and using deep learning process, and the climate prediction of adaptive control and solar power generation of wind generators in the energy industry It mentioned the optimization of power generation and reserve strategy forecasting.

Top

These changes are not only happening in certain industries, but they are happening in all industries, he said. But unfortunately, all companies are not involved in digital transformation.

The term digital transformation itself is to innovate the traditional industrial structure by applying technology, because it requires a lot of resources, costs, and time to make and internalize it.

He took an example of machine learning. Tasks such as detecting and classifying objects inside the image, or extracting metadata with OCR require machine learning-related technologies. However, there are more than 100 million business users who want to apply machine learning, while only 2 million actual machine learning experts and researchers. This inconsistency between such demand and supply causes the lack of manpower to develop and apply machine learning for business.

In 2012, the Andrew Naval Customization succeeded in classifying people and cat photos using 16,000 computers, but this simple work that simply distinguishes things requires a lot of resources, he said. Most companies are machine learning. Time and resources are not easy to invest in new investments.

Google Cloud offers Google's skills in the form of a public cloud, he said. Among them, AI and ML are provided in two models.

Google's AI and ML services first offer a 'pre-trained' model. This includes services such as translation, image awareness, and search of Google Services. There is no need to prepare a dataset for machine learning or create a model.

On the other hand, a custom model is provided. It is an industry-specific model that meets business requirements. You can develop everything yourself, but you can create customized models by putting your dataset in your machine learning model that you have learned in Google through auto ML.

Google Cloud's AI service is largely divided into visuals, language, dialogue, and structured data. In the visual section, it classifies images or videos, detects specific objects, and provides functions such as localization for location. The language part provides functions such as syntax analysis, emotional analysis, and translation for more than 50 languages, and the dialogue provides functions such as TTS and STT. The structured data part can be used for recommendation or prediction.

He then introduced 'Visual Inspection AI' as an industry-specific AI service. It is a service that allows AI to conduct inspections that have been visited by the naked eye in the manufacturing industry. Tasks such as cosmetic defects and PCB substrate assembly can be made in higher quality without relying on the test of examiners. Google Cloud's Visual Inspection AI provides features such as 'Animal Multisection', 'Detection and Finding', and 'Assembly Test'.

Visual Inspection AI provides active learning functions, so you can recommend Race on a data set without a label without a guidance learning that a person labeled and informs the answer, he said.


He demonstrated the process of performing assembly tests with visual instance AI services.

Park Hyung-jun, an engineer, said, The future business environment is gradually heading to a structure that is hard to survive without introducing cloud, big data, artificial intelligence, and IoT, and the dependence of information technology, IT technology in the business environment, is increasing. I hope you will actively participate in digital transformation with Google Cloud, the foundation of Google.

Comments

Popular Posts