| Case presentations at Google Cloud Next 2019
LG CNS AI Big Data Platform and | at the manufacturing site
| Google Cloud AutoML Application, Improved Defect Reading Time

LG CNS announced on the 11th that it is accelerating manufacturing intelligence by incorporating Google Cloud AI solutions.

On the 10th, LG CNS presented manufacturing AI innovation cases with Google at 'Google Cloud NEXT 2019,' held in San Francisco, USA, at the invitation of Google headquarters.

Google Cloud Next is a developer event hosted by the Google Cloud organization, and this year marks its fourth edition. It is becoming Google's largest event, with over 30,000 developers and partners attending this year.

Lee Seong-wook, Senior Vice President of AI and Big Data at LG CNS, who gave a presentation on this day, introduced a vision inspection case in which AI image recognition technology was applied to detect defects in manufacturing plant parts in collaboration with Google.

Since last year, LG CNS has been working to 혁신적으로 improve manufacturing defect rates by combining its AI big data platform 'DAP' with Google Cloud's AI solution 'AutoML (Automated Machine Learning).'

AutoML is a service that develops AI using AI; it is a technology that automatically generates new AI models applicable to specific fields by training AI with various data owned by a company, such as photos and text.

In the field of vision inspection, LG CNS DAP is responsible for the entire implementation of factory intelligence by analyzing tens of thousands of good and defective product data using AI deep learning. Google Cloud AutoML takes on the role of automatically generating an AI model for defect judgment by training the AI with the good and defective product data analyzed by LG CNS.

LG CNS solved the difficulties in the learning and modeling areas, which require significant time and effort throughout the entire vision inspection process, using Google Cloud AutoML.

LG CNS applied AutoML and tested it on hundreds of thousands of data points from LG affiliate manufacturing plants. As a result, it reduced the AI model creation period from one week to a maximum of two hours, recorded a reading time approximately 30 times faster, and increased the defect detection rate by about 6% on average. In addition, the combined technology of both companies has made it possible to achieve a reading rate of 99.9% even in processes with high judgment difficulty.

In addition, the introduction of AutoML has enabled developers previously responsible for AI training tasks to transition into data scientists and focus on highly specialized work.

LG CNS plans to accelerate the level of manufacturing intelligence by applying Google Cloud integration technology to improve defect detection and quality control in the manufacturing sectors of LG affiliates, such as LCD and OLED panels and chemical products.

LG CNS Executive Vice President Lee Seong-wook said, “We will leap forward as a leading global intelligent vision inspection company that provides vision inspection technology with 99.9% accuracy to smart factory sites.”