DX is an 'endless journey' | Ledge.ai
notebook-laptopLedge, which operates the AI-related media “Ledge.ai”, invites AI specialists from various companies and industries to discuss “issues and latest cases in each industry” and “technology including AI”. From this point of view, we regularly hold a web seminar "Ledge.ai Webinar" to discuss in a panel discussion format.
On October 26th (Tuesday), with the theme of "how to advance data utilization from 'idea' to 'execution'," Data Direct Networks Japan Co., Ltd. (hereinafter referred to as DDN) Representative Director Roberto A webinar was held with Mr. Triendl and Mr. Takeshi Izaki, General Manager of the Enterprise Business Headquarters of NVIDIA GK (hereafter NVIDIA) and Director of the Japan Deep Learning Association.
Digitalization is spreading in many companies, such as the abolition of faxes and the use of remote meetings due to the spread of COVID-19. However, there are not many cases where DX is promoted in terms of business transformation. In the first place, there are cases where the project does not move forward without knowing how to proceed with concreteness. In this webinar, we held a panel discussion on hints for companies to promote DX in the true sense, focusing on cases and trends regarding the fusion of business and technology. In this article, we will deliver the situation in the event report.
Speakers
Born in 1968 in Innsbruck, Austria. He graduated in 1992 with a master's degree in Mathematics and Music from the University of Innsbruck. After that, he studied abroad at the Research Center for Advanced Science and Technology of the University of Tokyo and the National School of Mining in Paris. After working at Mitsubishi Life Science Research Institute, became an independent consultant in 1998. After joining DataDirect Networks (DDN) in 2008, he became Senior Vice President of Global Markets in September 2016, overseeing all functions of DDN sales, marketing, product management, field service and support globally. there is He also serves as the representative director of DDN Japan.
After graduating from the Department of Materials Science, Faculty of Engineering, the University of Tokyo in 1997, completed the Department of Metallurgical Engineering, Graduate School of Engineering, the University of Tokyo in 1999. In 1999, he joined Texas Instruments Japan. After developing DVD application processors, mobile phone camera images, image signal processors, and DSP applications, he oversaw the digital product marketing department. Engaged in business development in a wide range of fields from entertainment products to industrial products. Joined NVIDIA in 2015 and worked as a business development manager for deep learning, and currently oversees the Enterprise Business Headquarters.
In my previous job, I worked as an editor for IT/digital web media and weekly magazines at a publishing company. She has also been involved in growth since the launch of paid membership media. Joined Ledge in November 2019. In July 2020, he became the editor-in-chief of Ledge.ai. In addition to writing articles, he has experience appearing in various media such as seminars and podcasts. I love unrivaled iced coffee.
The return on investment in AI has spread, but...
It has been a while since words such as AI, DX, and data utilization have spread. How has your approach to AI, DX, and data utilization changed over the past two or three years?
Mr. Izaki of NVIDIA says that AI (especially deep learning) is becoming more well-known, and the number of interesting cases is increasing.
――Izaki ``It's about deep learning, but our company started deep learning in 2015. However, at that time, even if we did walk-in sales, it was pretty much rejected (lol) After that, the media reported The awareness of deep learning has risen significantly.Although it is said that the spread is delayed, there are more and more interesting cases.One of them is Musashi Seimitsu Kogyo.Making car gears. It is a company that has a problem, but if there is even a 1mm defect, it is a big problem, so we had to do visual inspection to identify small defects.In the case of AI for this inspection, deep learning is used in the area of such anomaly detection. The number of cases where it is being used is increasing.”
DDN's Triendl says he feels it's only this year that he's been able to define the return on investment in AI based on his own experience.
--Trindl: I lived in Silicon Valley from 2017 to 2018 and visited various companies.Since then, I was wondering if it was possible to convert workstations to storage in order to handle larger amounts of data in Silicon Valley. After that, when I returned to Japan, I got the impression that Japanese companies are still experimenting with technologies such as AI at the workstation level.AI has to collect data, so In some cases, it costs more than the system.I have the impression that it was only this year that we were able to define the return on investment in such AI."
Two reasons why DX is not progressing in Japan
In the first place, is DX not progressing in Japan? If not, what is the reason? At the seminar, two ideas were presented.
Mr. Izaki said, "DX is not progressing in Japan," and pointed out two reasons.
--Izaki ``The progress is actually slow (laughs). There are several factors. This is a problem.In a sense, DX is similar to a new business.It is about creating business models and processes that create new value by linking data across multiple departments.In other words, we need leaders who can innovate. However, people who have had success with previous methods may become a force of resistance because they do not want to change the existing framework. For example, assigning people who are not successful in existing businesses, hiring excellent people from outside, etc.In the end, DX is an existing business There is also a place to destroy the business, so we need a completely new system.Regarding on-site engineers, it is quite difficult to recruit human resources from the outside, so we recommend that you first find them within the company. For example, some companies have discovered people who actually play with Python as a hobby on their days off, so it is important to give such people a free environment and discretion.”
The difference between Japanese and overseas DX is the in-house production of engineer resources?
On the other hand, Mr. Triendl thinks that Japan's DX is more advanced than Europe.
――Triendl: Considering my hometown of Austria, I can't say that DX isn't progressing in Japan (laughs).There are IT companies, the market is large, and there are many engineers. I think that there are many small and medium-sized companies in Japan, and it is a place where small and medium-sized companies are very active. There are high expectations not only for large companies but also for small and medium-sized companies."――Izaki: For example, Wal-Mart has over 1,000 engineers in-house and forecasts demand, but in Japan, in many cases, engineering resources are dependent on external vendors.In that situation, everything is manufactured in-house It is not realistic to standardize the system, so we first need to create a scheme that allows development with external vendors to work.At the same time, it is also important to develop internal human resources who understand that." --Trindl I think it's common both in Japan and overseas, but I think there is still a big difference in corporate culture.For example, in the United States, if a manager decides to do DX, it will proceed quickly.Europe is rather similar to Japan. Kana ....When it comes to AI in DX, AI is not just about introducing an application, it is more like research.It is necessary to evaluate and proceed with the model that is created while trying it little by little.So companies And university collaboration is also important.What is noticeable in the United States is that early AI projects adopt entire university departments.Uber adopted the entire robotics department of Carnegie Mellon University and established an autonomous driving department. I think there is a good relationship between universities and companies, as I have mentioned.”
No matter how good a manager or engineer you are, DX takes time
The word DX has been around for a long time, but at the same time, it tends to be used without a clear definition. Ultimately, what can be said to be DX?
Izaki points out that DX is a “never-ending journey.”
――Izaki ``We call it DX, but before DX, there is digitization, which is the transition from analog to digital, such as converting stamps into electronic signatures, and then adding value using digital. There is digitalization to create.DX is to make that value creation perpetually rotate, and it is necessary to create an organization, create a culture, and continue.Creating a culture is endless, isn't it? I see DX as a journey that never ends."Triendl agrees that DX is a cultural change that takes time.
――Triendl: As Mr. Izaki says, DX is not a single project, so I think it's a correct expression to say that it's a never-ending journey. I see it as a combination of things.For example, 30 years ago, research aimed at making it possible to simulate crash tests in a computer to measure the safety of cars was said to be a wasteful investment at the time. However, today it is nothing less than having data that can be simulated on a computer, and the data are becoming more closely linked.
DX means handling these data, and it started more than 10 years ago. Even if the data is not connected now, it will be connected in 10 or 20 years. In that sense, I think it's correct to say that DX is a culture."
At the seminar, the audience asked the following questions.
Q. I think that the resolution of the management team is ultimately necessary for promoting DX, but on the other hand, even if people have ideas, they are so busy with their daily work that they have no time to think about DX. Who should start in this case? --Izaki "That's a question that makes me feel compelled. As you say, it's the management who can make the proper investments. It is impossible to do it on site alone, so management must be prepared.On the other hand, on the other hand, the site should study deep learning and other technologies even if there is no time, and In order to be selected when the wave of DX comes, we should raise our technical presence within the company.”
――Triendl: I think culture is important for change. It is the same for an IT company like ours, but even in the work of supporting customers' IT, the decision whether to discard or utilize data is important. It is important to have a culture that becomes.Because long-term investment is absolutely necessary.DX is a change related to corporate culture, so it takes time, no matter how talented engineers are.In that sense, Mr. Izaki's DX It feels right to travel.It's important not only for the management, but also for the company as a whole." , What should a company that is finally starting DX start with? What the two companies had in common was the importance of first identifying the problem.--Izaki ``The most important thing is to identify the issues first, rather than blindly calling it DX. We have also prepared a program called the Application Program, which allows you to match with excellent external partners.On the other hand, the complete division of labor between the external and the internal is not a good idea, for example, when optimizing the application for the hardware. It depends, but only the application knows about this kind of thing.It is better to understand the information properly and decide whether to outsource it or do it in-house.But before that, Again, it starts with identifying the problem.”
――Triendl ``As Mr. Izaki said, DX is not first, but business issues are first. We want to provide the infrastructure to solve these problems." About DDNData Direct Networks (DDN) is a high-performance computing ( HPC) is a company that has accumulated many achievements in storage systems. Over the past 20 years, it has been installed by more than 10,000 customers, including enterprises, government agencies, and the public sector, and has 10 data centers globally. Furthermore, in recent years, we have acquired companies that provide storage systems for enterprises one after another, software-defined storage (SDS) Nexenta's multi-cloud data management function, IntelliFlash's edge computing function, and Tintri's virtual infrastructure storage function. We offer a wide range of products and support services as a total provider of storage systems.Data Direct Networks Japan Co., Ltd. (DDN Japan) was established in 2008 as a Japanese corporation of DDN and a wholly owned subsidiary. We have a large-scale benchmark lab in Tokyo, and provide numerous benchmark tests and PoCs to domestic and overseas customers. In addition to hardware sales, we have an engineering team that can provide support in Japan.In the past seven years, we have provided domestic customers with extremely large-scale storage systems that have reached 1.4 exabytes. is also accelerating.
◆DDN's AI storage has been verified on a large scale with NVIDIA SuperPOD, accelerating deployment with a number of advanced AI platformsDDN's AI storage has been verified with NVIDIA SuperPOD and is NVIDIA's AI supercomputer Adopted as storage for "Selene". Selene is an extremely high-performance AI supercomputer ranked No.6 in the June 2021 Top500 list, but by adopting 40 DDN AI400X appliances, it can achieve a maximum throughput of 2TB/sec and 120 million IOPS. shows very high performance. The results of large-scale verification with DDN storage on SuperPOD are provided by NVIDIA and DDN as reference architectures or blueprints.
Document download: https://ddn.co.jp/ddn_channel/resource-library.html
◆Information about GTC [November 8th (Monday) to 11th (Thursday)]For 4 days from November 8th (Monday) to 11th (Thursday), numerous sessions and technical training will be held, including the latest information on AI and deep learning, and success stories in each country. . In addition to more than 500 sessions and training workshops by NVIDIA experts, world-renowned researchers, and Japanese business leaders who are driving DX (Digital Transformation), the event is full of inspiration and appeal. , We also have a place for exchanges. Please join us.
◆ Participation fee is free | Pre-registration URL: https://nvda.ws/3AclkaP
On Tuesday, November 9th at 17:00 GMT, NVIDIA Founder/CEO Jensen Huang will give a keynote address. There are many surprising new proposals that solve the difficult problems facing the world through the development of technology. On November 10th (Wednesday) at 11:00 (Japan time), there will also be a featured session in Japanese by Mr. Otsuka, president of Musashi Seimitsu Industry, who appeared in this interview, and Mr. Izaki of NVIDIA. (Session ID: A31613)In order to explore the differences between companies that have succeeded in DX and those that have not, we will approach the "organizational image" and "manager image" of successful DX companies through an interview with President Otsuka.
