Overview

GMO Internet, Inc. (Chairman, President and Group CEO: Masatoshi Kumagai, hereafter "GMO Internet") and AVILEN Corporation (hereafter "AVILEN"), a group of AI specialists who provide comprehensive support for AI businesses, have signed a collaboration agreement to promote the use of AI technology using GPU (Graphics Processing Unit) computing platforms in Japan. (AVILEN), a group of AI specialists who provide comprehensive support for AI businesses, have signed a collaboration agreement to promote the use of AI technology using GPU (Graphics Processing Unit) computing platforms for business applications in Japan.

Based on this agreement, from today, October 3, 2019 (Thursday), the "GNA Program" () will be offered to companies seeking to utilize AI technology, providing support for GPU computing platform environments, consulting on specific applications in business operations, businesses, and services, and AI human resource development. (https://gpu.cloud.gmo.jp/gna/). In addition, the two companies will jointly hold seminars and technical exchange events that leverage AVILEN's knowledge, and work to develop an environment for fostering AI-savvy human resources, with the aim of raising the level of AI business in Japan.

Background of the Collaboration

According to a survey by a U.S. consulting firm, companies that introduce AI into their business by 2025 will increase their revenue by approximately 120% in 2030 compared to 2017, while companies that do not introduce AI into their business may decrease their revenue by approximately 20% in the same year (*1). (*1).

Thus, the introduction of AI has been recognized as a demanding element in driving business growth, and its adoption has become an urgent necessity on a global level. However, in Japan, there are few examples of AI applications in various industries, making it difficult to understand its cost-effectiveness (*2), and the supply of human resources familiar with AI and other advanced technologies has not kept pace (*3).

GMO Internet, which provides GPU computing services for machine learning and deep learning, and AVILEN, which provides comprehensive support for AI businesses through the development of AI system models and human resource training, will leverage their respective areas of expertise and work together to develop a joint project to promote the use of deep learning, machine learning, and other AI technologies in Japan. The two companies have decided to collaborate in order to promote the use of AI technologies such as deep running and machine learning for business in Japan by taking advantage of their respective areas of expertise.

(*1) McKinsey & Company "Notes from the AI frontier: Modeling the impact of AI on the world economy" https://www.mckinsey.com/featured-insights/ artificial-intelligence/notes-from-the-ai-frontier-modeling-the-impact-of-ai-on-the-world-economy
(*2)Ministry of Finance, "The State of Utilization of Advanced Technology in the Survey of the Bureau of Finance," https://www.mof.go.jp/about_mof/zaimu/kannai/201803/sentangizyutuzirei091.pdf
(*3)Ministry of Economy, Trade and Industry, "AI Human Resource Development Initiatives"
https://www.meti.go.jp/shingikai/economy/jinzai_ikusei/pdf/001_03_00.pdf

Overview of the GNA Program

The "GNA Program" is a program that supports the value enhancement of customer services and business growth by promoting the use of AI in terms of both hardware and software, such as the development of a GPU computing environment and support for the introduction of AI technology into business operations.

Specifically, GMO Internet, which has one of the largest Internet infrastructure platforms in Japan with more than 10 million customers, will provide the GPU computing platform "GPU Cloud by GMO" for machine learning and deep learning, as well as support for AI use environment. In addition, AVILEN, a company that operates both AI human resource development and AI development businesses, will provide consulting services for specific applications and AI human resource development training services that incorporate the wealth of knowledge from the learning media "Statistics for All Mankind.

Name GNA Program
Support details ・Provision of "GPU Cloud by GMO" at a special price for a certain period (1 month to 1 year)

• AI implementation consulting
• Introduction of partner companies and use cases of other companies necessary to solve problems
• Technical advice, software environment selection, and environment construction support
• Technical training (provision of AI human resource development training services for customers of this program of AVILEN's "Statistics for All Human Beings")

*Support contents may vary depending on the stage of consideration in the introduction of AI/GPU utilization by the client. *There are some conditions for participation in this program. Please refer to the following website for details.

Program Details https://gpu.cloud.gmo.jp/gna/

About Free Courses of "GNA Program" AI Human Resource Development Training Service

AVILEN's AI training service, with a total of more than 10,000 trainees, offers more than 100 different courses covering the basic level for beginners to the advanced level for experts, from zero to the practical level. The courses are based on a wealth of knowledge by management team members with specialized research and practical experience in AI, and include many case studies and other exercises to enhance practical skills. We are also one of nine companies (*4) with educational courses approved by the Japan Deep Learning Association (hereinafter referred to as JDLA), and we offer courses corresponding to the G certification and E qualification (*5).

AVILEN, which has both AI human resource development and AI development businesses, has prepared three courses for "GPU Cloud by GMO" customers, which are specialized training courses that incorporate knowledge gained from the development field and can only be realized by AVILEN. The content of the courses will be updated as needed, and they will be available free of charge to "GPU Cloud byGMO" customers. For specific course content, please refer to the [Attachment (reference material)].

(*4)As of October 1, 2019
(*5)G certification is a qualification test for AI generalists conducted by JDLA; E certification is a qualification test for AI engineers certified by JDLA.

Comment from Masataka Osaki, NVIDIA Japan Representative and Vice President of U.S. Headquarters

NVIDIA is pleased to welcome GMO Internet, Inc. to the GPU Cloud by GMO and the GNA Program, NVIDIA's GPU computing platform is a high-speed, high-performance computing platform for learning and operating deep learning. We look forward to NVIDIA's cutting-edge GPUs and software offerings contributing to the diversification of GMO Internet's services and the further revitalization of the AI market.

Kotaro Takahashi, Director of AVILEN Corporation

AVILEN Corporation sincerely welcomes the release of the GNA Program. We expect further development of the AI market as more and more people work with GPUs that enable high-speed processing. In addition to the development of AI-based systems, we provide comprehensive support for the AI business by training personnel to handle these systems. We will continue to cooperate with GMO Internet Inc. and NVIDIA to promote support for customers suffering from AI throughout Japan.

Future Development

GMO Internet, AVILEN, and NVIDIA will collaborate to provide products that facilitate the commercialization of AI and machine learning, and to train human resources to promote such products. By doing so, the three companies aim to raise the level of AI business in Japan and increase the value of their businesses and services.

Future initiatives with NVIDIA

GPU Cloud by GMO" is currently undergoing verification so that it can take advantage of NVIDIA's "NGC," a software hub covering AI, data science, HPC, etc. Servers equipped with NVIDIA GPUs can be used in data centers, clouds, edges, etc. to build programming environments for AI implementation. NVIDIA's service is expected to lower the hurdles for utilizing AI technology by reducing the time and effort required to build and set up a programming environment when introducing AI. The enhancement of the "GPU Cloud by GMO" service through the use of NVIDIA's services is scheduled to begin by December 2019.

Future Initiatives with AVILEN

We will promote the use of "GPU Cloud byGMO" for AVILEN's business AI application support, and will utilize the knowledge and case studies gained through the verification and utilization of AI to improve and enhance "GPU Cloud byGMO". This is expected to lead to the provision of more flexible and secure GPU computing resources.

What is GPU Cloud by GMO?

NVIDIA GPUs are processors (computing devices) developed for image processing (rendering) such as 3D graphics and video, which require advanced parallel processing, and are now increasingly in demand in situations that require advanced computational processing such as machine learning, deep running, and AI.

GPU Cloud by GMO" uses NVIDIA's latest GPU "NVIDIA® Tesla® V100" and combines high-speed computing resources with high-speed interconnect "InfiniBand EDR" and distributed file system "Lustre". This service is provided via the cloud as a bare-metal server (*6). Customers do not need to invest in their own facilities, and can procure computing (GPU computing) resources when they need them and as much as they want to use them.

The "NVIDIA® Tesla® V100" used in "GPU Cloud by GMO" delivers performance equivalent to up to 100 CPUs per unit (*7), and can provide powerful support to engineers and researchers tackling issues that require massive data processing, which has been considered difficult until now. (*6) A physical server.

(*6) Physical server.
(*7) See NVIDIA® Tesla® V100: https://www.nvidia.com/ja-jp/data-center/tesla-v100/

AVILEN Corporation (URL: https://avilen.co.jp/)

Company name AVILEN Corporation
Address: 6F Rivera Iwamotocho, 3-5-14 Iwamotocho, Chiyoda-ku, Tokyo, Japan
Representative: Choi I-Ming
Description of business 

• AI human resource development business
• AI business consulting and technology development business

Capital 3 million yen

GMO Internet, Inc.（URL：https://www.gmo.jp/ ）

Company name: GMO Internet, Inc.
Location Cerulean Tower, 26-1 Sakuragaoka-cho, Shibuya-ku, Tokyo, Japan
Representative: Masatoshi Kumagai, Chairman, President and Group Representative
Business Description

• Internet Infrastructure Business
• Internet Advertising and Media Business
• Internet Finance Business
• Virtual currency business

Capital 5 billion yen
Copyright (C) 2019 GMO Internet, Inc.
Attachment (Reference Material)

Courses covered by the GNA Program AI Human Resource Development Training Service

Introduction to Python (*8)

This course provides a hands-on introduction to Python grammar so that even programming beginners can understand it. This course aims to help students learn Python quickly through a hands-on format with exercises. The course is available for video review for one week from the day after the course date. Learn basic Python syntax.

Able to write simple programs in Python and understand Python code while examining it

Introduction to Python Machine Learning

This course aims to cover the whole picture of machine learning through lectures on scikit-learn (hereafter, sklearn), a machine learning library, and to enable students to implement machine learning on their own. The course is open to the public for one week from the day after the course date for video review. Learn how to use sklearn

Experience implementing machine learning methods

Understanding of appropriate application scenarios for each machine learning method・Mastering of appropriate machine learning approaches using sklearn

This course explains the mechanism of CNN (Convolutional Neural Network) (*9) and provides an implementation using tensorflow. This course is intended for those who understand the basic syntax of Python3 and have experience building simple neural networks (multilayer perceptrons) with sensorflow. ・Acquisition of basic principles and implementation methods of deep learning and CNN ・Gain a bird's eye view of what can be done with CNNs.



(*8)Python is a programming language with an easy-to-understand and simple syntax that is used in many libraries required for utilizing AI technologies, such as sensorflow.

(*9)CNN is a neural network that offers superior performance, especially in the field of image recognition.