Newswise — Since 2021, Korean researchers have provided a major boost by providing a simple software development framework to users with relatively limited AI expertise in industrial sectors such as factories, healthcare, and shipbuilding.
ETRI announced that it has released the core technology of its MLOps tool, which automatically generates neural networks based on no-code and automates the deployment process, as open source on GitHub.
On November 6th, the research team held the 4th public seminar to expand the TANGO GitHub community at the Science and Technology Center in Gangnam-gu, Seoul.
The TANGO framework is a technology that automatically develops application software (SW) that applies artificial intelligence and optimally deploys it to various target hardware (HW) environments such as cloud, Kubernetes on-premises environment, and on-device.
For example, in quality inspection at a steel mill, it is easy to determine whether steel data has defects, but applying AI has not been easy. At hospitals, it is easy for doctors to diagnose tuberculosis just by looking at X-ray images, but it has been difficult to utilize automatic prediction models using AI.
The TANGO framework developed by ETRI is well suited for neural network processing tasks for domain experts who lack extensive neural network knowledge. It's also easy to use, with a simple installation command that installs automatically and runs right from the web interface.
In traditional AI application software development methods, domain experts are responsible for labeling data, and software developers are responsible for developing and learning artificial intelligence models, and implementing and running application software.
However, with the expansion of artificial intelligence (AI) technology, the demand for software (SW) is increasing in all industries. On the other hand, there is a shortage of AI and SW specialists who can meet this demand.
To address these issues, ETRI has developed and officially released a neural network automation algorithm optimized for object recognition, reflecting the needs of domestic industrial sites. ETRI also released the LLMOps tool to support the development of generative AI.
During the development of TANGO, 24 domestic and foreign patents were invented, 3 NeurIPS papers and 13 SCI papers were published, 4 technology transfers and commercialization income of 10 billion won were brought.
Autonomous navigation solution company Abenotix was selected for the “Public Research Results Expansion and Commercialization Project'' supported by the Ministry of Science, Information and Communication, which aims to commercialize excellent research results from government-funded research institutes through tango technology transfer, and received investment of 1.3 billion won (corporate value: 9.8 billion won) from Korea Science and Technology Holdings (500 million won), Korea Credit Guarantee Fund (500 million won), and Korea Science and Technology Holdings. Low Partners (300 million won).
Through technology transfer, Abenotix has secured Tango's on-device deployment technology and AI performance optimization technology, and is commercializing on-device AI that automatically generates the context information needed by navigators.
The research team is focused on broader adoption by conducting pilot demonstrations led by partner research institutions.
Weda Co., Ltd., a joint research institute, has developed an artificial intelligence service that can be used by on-site employees at two companies in the steel and automobile parts manufacturing fields. Specifically, this service was developed for vision-based visual inspection of more complex geometries, such as automotive bumper rolls.
Lablup Co., Ltd., a joint research institute, has collaborated with KT Cloud to launch an implementation optimization service that supports Rebellion's latest domestic AI acceleration engine “ATOM-Max.” We have also commercialized a GPU cloud rental service in collaboration with KT.
Seoul National University Hospital is developing artificial intelligence technology that uses large-scale chest CT images and diagnostic data to automatically generate diagnostic reports from CT images. The developed technology will be tested to provide a cardiopulmonary disease prediction service at four hospitals including Seoul National University Hospital (Seoul National University Hospital, Seoul National University Bundang Hospital, Seoul National University Hospital Gangnam Center, and Boramae Hospital), and will be evaluated and verified using actual clinical data.
In particular, the LLMOps tool, which supports the development of generative AI, is being developed in collaboration with Acrylic, with a view to immediate commercialization. The source code for Acrylic's commercial product, Jonathan, is fully publicly available on GitHub, with the addition of core algorithms and the creation of a standard operating environment for industry-specific generative AI applications.
Kim Tae-ho, software PM of the Institute of Information and Communication Technology Planning and Evaluation (IITP), said, “TANGO technology is truly Korea's best open source project, and it is greatly contributing to strengthening the competitiveness of the domestic software industry in the field of artificial intelligence development tools.''
ETRI Chief Researcher Cho Chang-sik said, “We plan to expand the existing Tango project, which utilizes vision neural networks, to the field of LLMOps tools that support generative AI.In the future, we will share all our development know-how and provide solutions that can be directly commercialized by the industry through verification.''
The research team said it will continue to release new versions of the source code on GitHub every six months. In addition, we plan to hold a public seminar once a year in the second half of the year to share not only the source code for technology development but also practical know-how. Meanwhile, a total of 944 people from 552 institutions attended four sessions at the Tango public seminar and shared their insights on AI technologies.
1) Machine Learning Operations (MLOps): MLOps stands for Machine Learning Operations and is a technology and tool for managing the machine learning lifecycle, including data preprocessing, model development, deployment, and operation.
2) No-code: A development approach that enables faster and more accurate application development with an easy-to-use interface, even for users with limited coding experience.
3) GitHub address: https://github.com/ML-TANGO/TANGO
4) TANGO: TANGO (Target Aware No-code Neural Network Generation and Operation Framework)
5) Kubernetes on-premises environment: This refers to an environment where factories, hospitals, etc. operate their own servers and data centers instead of using a cloud environment for security reasons. This means deploying and managing Kubernetes on your own physical infrastructure (servers, networks, etc.) without using an external service provider's infrastructure. Kubernetes is an open source system for deploying and managing containerized applications.
6) Data labeling: The process of identifying and annotating data to train machine learning or artificial intelligence (AI) models. Enable learning models to understand data and make predictions. For example, to train an image recognition model, descriptions (labels) of objects (cats, dogs, cars, etc.) in photos are added.
7) LLMOps: LLMOps is an abbreviation for LLM Operations, which is a technology and tool for managing the life cycle of large-scale language model (LLM) training, including data preparation, training, tuning, deployment, and operations.
###
This result was developed with support from the Ministry of Science, Technology and Information Communication and the Institute for Information and Communication Technology Planning and Evaluation (IITP) through the “Automatic generation of neural network applications and optimization of execution environment'' and “Generation AI support system SW framework'' projects.
About Electronics and Telecommunications Research Institute (ETRI)
ETRI is a government-funded, nonprofit research organization. Since its establishment in 1976, ETRI, a global ICT research institute, has made great efforts to bring about remarkable growth in Korea's ICT industry sector. ETRI positions Korea as one of the world's top ICT nations by continuously developing the world's first and best technologies.
