BUSINESS

ADAMS Platform

AI Vision technology helps improve the quality of factory automation without worker intervention, and make sophisticated decision. The Data that needs for learning and decision making of AI must have high accuracy and reliability.
ADAMS, LAONPEOPLE AI platform, can maximize both AI utilization and cost-effectiveness for factory automation by providing systematic management of data and a flexible structure so that AI can make the right decisions.

Why LAON PEOPLE
LAONPEOPLE developed the ADMAS solution, based on accumulated know-how of using various product data and AI inspection experience in industrial field for many years. ADAMS is optimized for AI inspection and data utilization in whole process including gathering, analysis and management of data.
INQUIRY
sales@laonpeople.com031-707-7052ONLINE INQUIRY
LIST

Product Specific Specifications

Information
SCROLL >>

 

ADAMS

AI Data Analysis Management System

ADAMS is an AI data management platform that composed with five tools.
It is a MSA (Micro Service Architecture)-based platform service that provides
essential functions for maintaining optimal AI performance, data management, and collaboration.

 







I thought it would be the done if I apply AI...

ADAMS can solve various problems that occur in industrial fields.



Data gathering stage Data refine(labeling) stage Model learn stage Model evaluation stage
* Need to develop gathering process
* Impossible to generalize the gathering process
* Unable to trace gathering process		 * Unable to share labeling work
* Unable to distribute work
* Unable to trace work process		 * Unable to configure data set freely
* Unable to automate learning		 * Unable to automate or generalize scheduling
EDEN: Data Management Tool Information
Eden provides file storage and DB functions as a data lake.





Centralization of data
    Eliminating data redundancy by data centralization. It also helps data management by demonstrating the integrity of the data.

Supports rest API
    Rest API is a general-purpose web API that does not limited by programming language and space.
    Therefore, easy integration is available when developing customer applications.
TREE : 라벨링 협업 Tool Information
Tree provides labeling and scheduling function.




Task management
    Administrator assign the tasks to operators which require manual labeling, and the operator completes the assigned task by the due date.
    During this process, whole task histories are logged to visualize statistics. This gives administrators a detailed view of the task.

Semi-Auto Labeling
    Tree enables faster labeling and helps to improve productivity, by providing a labeling UX that is easy to work with.
EVE : 모델 학습 및 테스트 Tool Information
Eve provides functions for learning/inference.



Supports open framework.
    Supporting Pytorch / Tensorflow / Keras allows users to operate ADAMS platform with any AI engine they want.

Visualize data
    EVE helps users to gain insight by visualizing learning results / inference results / test results.
SNAKE : 지속 학습 및 배포 Tool Information
Snake is a tool that provides continuous learning of models.



Semi-auto learning
    AI models require continuous learning. This is because the performance of the AI model itself is constant, but the data in reality is constantly changing.
    A lot of resources should be put into this process. Snake improves performance and reduce resource by automating continuous iterative learning.

Model update
    An agile respond to changes in industrial fields can be available by automating the latest AI model obtained through auto-learning.
UNIVERSE : 데이터 프이프라인 관리 Tool Information
Universe is a tool to manage the data pipeline.




Data Input pipeline management
    Data accumulated from industrial fields is stored in online storage according to specific criteria which managed by ADAMS.
    Data is automatically uploaded to the appropriate storage classified by context of data.

Image processing pipeline management
    For accurate vision inspection, the first acquired image is changed in various ways to acquire more suitable images for the inspection.
    This image processing process can be configured by the user freely and provided in a reusable form, making it easy to acquire the image processing process.

Learning Pipeline Management
    Systematize the entire process required for learning, from data input, labeling, setting, learning, testing, and performance verification,
    to automate each step of the task. This helps to identify and resolve the problem accurately, by monitoring all the process.
ADAMS 적용사례 Information
Smart Factory



    According to situation that production field data changes continuously, the data needs to be relearned in a short period continuously.
    The automatic learning function of the Snake package replaced many tasks that had to be done every time and reduced the manpower and time required for this.