An Exclusive Interview with Dr. Seishi Okamoto, Fellow of Fujitsu Research, Fujitsu

Here is an exclusive interview with Dr with Dr. Seishi Okamoto, Fellow of Fujitsu Research, Fujitsu

The future of information and communication technology is growing faster than expected and it is time to think about its sustainability. Through innovation, tech companies are building extensive digital designs that would be sustainable as well as adaptable in the digital world. As the leading Japanese information and communication technology (ICT) company, Fujitsu offers a full range of technology products, solutions, and services. It has extensive experience in designing, building, and deploying IT systems, services, and digital solutions for both public and private sector customers from retail and healthcare to financial services and automotive.

Analytics Insight has engaged in an exclusive interview with Dr. Seishi Okamoto, Fellow of Fujitsu Research, Fujitsu.

1. Brief us about the company, its specialization, and the services that your company offers.

Fujitsu's purpose is to make the world more sustainable by building trust in society through innovation. Our aim is to build new possibilities by connecting people, technology, and ideas, creating a more sustainable world where anyone can pursue their dreams.

As the leading Japanese information and communication technology (ICT) company, Fujitsu offers a full range of technology products, solutions, and services. We have extensive experience in designing, building, and deploying IT systems, services, and digital solutions for both public and private sector customers from retail and healthcare to financial services and automotive. Fujitsu's major technological contributions include the world-leading supercomputers "K Computer" and "Fugaku" (jointly developed by RIKEN and Fujitsu). Since being founded in Japan in 1935, we have continued to demonstrate our technological prowess and the persistent pursuit of innovation. As a world-leading digital transformation partner, our business structure is aligned with the modern digital world.

Using a wide portfolio of trusted technology services, solutions, and products, we work with our customers to co-create solutions that help them on their journey to enterprise-wide digitalization.

At the same time, we are using technology and working with our customers and wider ecosystem partners to help solve social issues. By making a contribution to the United Nations Sustainable Development Goals, we are helping to transform our world and build an inclusive, sustainable, and trusted society.

2. How is big data/AI evolving today in the industry as a whole?

We believe that organizations are today encountering unparalleled levels of complexity, whether that's pandemics, sprawling globalized supply chains, rising customer expectations, or ever-increasing digital demands. There is huge value in data and understandably a drive towards digital transformation initiated in every organization. More enterprises than ever are assessing the opportunities hidden in their treasure troves of data to supercharge their business and take the lead in their field.

To succeed and grow in this challenging world, businesses must adapt at speed and find new ways of working to generate ongoing business value.

However, businesses can only be as successful as the data they capture and integrate into their enterprise operations, which is why applying data and analytics at every opportunity is fast becoming a prerequisite for organizational success and disruptive decision making. Owing to the sheer volume, velocity, and variety of data available, AI is proving instrumental in driving business insights, minimizing the probability of human error, improving overall business efficiency, and increasing innovation in today's connected world.

In some cases, AI can analyze datasets in an hour that would've taken researchers many years, making it a strategic technology to tackle the most pressing of challenges confronting humanity.

3. What are the key trends driving the growth in big data/AI?

At Fujitsu, people remain central to our vision, with AI solutions centered firmly on empowering workers and citizens, creating value, and supporting the work they do. The impact of AI on our lives may not be something that we will notice or appreciate overnight, however, one thing is clear – its progress is constant and inevitable.

With its ability to identify patterns and detect anomalies in mountains of digital information, AI is opening new possibilities in a range of disciplines and areas of business; once trained, it's tireless in processing many standard tasks. For example – by adding AI to service desks and call centers, staff are freed from performing low-level, monotonous tasks, and are enabled to focus on addressing more complex technical problems/complicated requests or delivering an enhanced customer experience/improved customer care. In this way, in the business world, AI is radically transforming industries like manufacturing, healthcare, transportation, retail and financial services.

We believe that in the future, there will be an even stronger need for AI technologies that allow users to discover things unseen by humans and to support more advanced decision-making. Thus, a "Trusted AI", that is "explainable," "understandable," and "ethical" will be the trend.

4. What is the edge of your company that differentiates it over other players in the industry?

Fujitsu 's goal is to offer transparent, ethical, and accountable AI technologies customers can trust. We have customers globally and many kinds of technologies to satisfy customers' requirements.

At Fujitsu, we have been creating AI solutions for more than three decades and commercialized Japan's first computer with AI, FACOM Alpha, in 1985. In 2015, Fujitsu launched the Fujitsu AI Zinrai Platform. Zinrai AI provides a human-centric approach to the co-creation of services and solutions. It utilizes 'best of breed' global technologies that have been developed and deployed to meet ever-growing customer challenges. "Zinrai" has been used in diverse areas ranging from cancer genomic medicine to safe navigation for ships.

"eXplainable AI (XAI)":

Making AI and its decisions explainable is one of the most important tasks in many industries as the concept of "eXplainable AI (XAI)" shows.

Deep learning systems, with their complex neural networks inspired by the human brain, are often like "black boxes." Such systems also require a large volume of training data to make accurate predictions.

To solve such issues, Fujitsu included the "Wide Learning™" machine learning technology as a core component in its Human Centric AI Zinrai. Wide Learning™ is an AI technology that implements the process of scientific discovery on a computer by reproducing the logical and objective thought process in science. This technology evaluates every combination of data items in a system to find important combinations, and thus enables AI to quickly test a vast number of hypotheses that humans have been unable to find. It can also use the discovered hypotheses to automatically make intelligent decisions such as predictions and classifications. In contrast to deep learning systems, the Wide Learning™ approach can generate accurate predictions even with a small volume of training data.

One example of a practical application of Fujitsu's Wide Learning™ technology is the functional analysis of COVID-19 policies with the aim to discover important combinations of characteristics and policies that are related to the limitation of infections in different countries and regions. Moreover, Fujitsu's Wide Learning™ has been applied during the 49^th House of Representatives elections in Japan to analyze and explain the features of the candidates who win or lose elections.

"Discovery AI":

Fujitsu further contributes to the healthcare sector through cancer treatment technology.

We already have a piece of extensive knowledge about the causes of cancer stemming from gene disorders. However, developing new treatments is a slow and painstaking process, involving checking relationships between over 20,000 different genes and cancer variants one by one.

Rapidly advancing AI technologies will hold the key to resolving this issue. If we can achieve a paradigm shift by applying AI to cancer and gene analysis, we can open up a completely new pathway for personalized gene therapy. This could lead to a dramatic improvement in the success rate for treating cancer and saving even more lives.

There are two important areas we need to address before we can securely tap into AI's potential for cancer treatment. The first one is "accountability" for AI, ensuring that AI can clearly explain how a judgment is made in ways clinicians can understand. The other one involves the "inference of causalities," meaning that AI can clarify the reasons why specific genetic mutations cause cancer.

Fujitsu is tackling challenges to achieve these two breakthroughs. "Deep Tensor," Fujitsu's Explainable AI technology, allows us to very accurately analyze graphical-structural data, which can represent the complex connections between people and physical elements. Deep Tensor analyzes the entirety of huge gene networks in minute detail, suggesting which gene mutation and associated factors are likely to trigger cancer progression. This is further supported by medical knowledge drawn from existing scholarly articles, helping doctors use the analyses provided by AI with confidence.

Thanks to the power of "Fugaku," the world's fastest supercomputer developed jointly by Fujitsu and RIKEN, these types of calculations can be completed within a single day, compared to months using a standard supercomputer. As a result, using AI to research new cures for cancer has rapidly advanced into the practical sphere.

At Fujitsu, we are aspiring to go well beyond this, applying our expertise to realize "Discovery AI". Normal machine learning principally reveals correlations between the data, rather than causalities, which involve the relationship between cause and effect. If we are to identify the real causes of cancer and establish effective new cures, we have to pin down these causalities.

Fujitsu is conducting R&D to realize Discovery AI that can reveal causalities among huge amounts of data from the human genome and lifestyles, offering a reliable hypothesis of the most effective treatment. We have already applied this approach to calibrate a causal relationship between different types of colorectal cancer and approximately 300 related genes. From the data only, we succeeded in rediscovering the genes related to a particular type of colorectal cancer. Our next step is to work on the discovery of causalities based on the approximately 20,000 genes that compose the human genome through our AI technologies. In this way, we will continue to contribute to the advancement of cancer treatments by enabling creative collaboration between medical professionals and AI technology. Our fundamental goal is to contribute to solving difficult social challenges and improving the wellbeing of people. Fujitsu will continue to focus on technology-based innovations that will enable previously impossible discoveries and help people live healthier, more secure lives.

"Trusted AI":

Another ground-breaking technology of Fujitsu is "High Durability Learning".

While the number of companies using artificial intelligence (AI) for their business is increasing, the "deterioration of AI model accuracy during long-term system operation" has emerged as a new challenge. To solve this issue, Fujitsu developed "High Durability Learning". This world-first technology can estimate the accuracy of an AI model at any time during its operation and automatically restore it when deterioration of its accuracy is detected. This technology thus enables AI models to be maintained with high accuracy over a long period and enables stable AI operations in various circumstances.

This technology has, for example, been applied to the financial sector during a verification test using financial data from 3,800 companies, and to the smart city, sector to recover AI accuracy during the operation phase due to degradation.

Furthermore, the realization of an" ethical" AI is an urgent task in our current society, as the many news about unethical, discriminating AI shows. Fujitsu is one of the pioneers in the area of "Ethical AI", and we have been collaborating with AI4People, Europe's largest forum for AI-related ethics. In 2019, Fujitsu formulated the "Fujitsu Group AI Commitment," demonstrating its dedication to the safe and secure use of AI. All Fujitsu's AI-related activities are aligned with this commitment to providing our customers with an AI they can trust.

"AI and Fugaku":

"Fugaku," the world's fastest supercomputer, jointly developed by Fujitsu and RIKEN, can perform difficult and time-consuming calculations that would require months to complete with a standard computer within a single day. Fujitsu is usingFugaku for unique AI applications, including a simulation application to predict tsunami flooding that helps to ensure peoples' safety and security, developed in cooperation with Kawasaki city. We also conducted actual evacuation drills and verified the feasibility and usefulness of our technology in the real world.

Fugakufurther took first place in the MLPerfHPC benchmark, whichdemonstratesFugaku's superiority not only as a classical supercomputer but also as an"AI supercomputer".

5. Brief us about your role at Fujitsu and your journey in this highly promising sector.

I started my career at Fujitsu 30 years ago, in 1991, at what was then called"Fujitsu Laboratories," following the call of the then Head of Research. The reason why I joined Fujitsu was that I wanted to do AI research to develop AI products and technologies that add value to our society. Back then, Fujitsu was one of the very few companies that had a specific department that focused on AI research. After joining Fujitsu, I first spent 12 years as an AI researcher, with a special focus on case-based reasoning, machine learning, language processing, and information retrieval. After this, I spent eight years as a manager at the AI research Department (Associate Researcher, Director of Research). In 2011, I was transferred to Fujitsu Limited, where I was leading the data science business to develop Fujitsu's big data business. Here, I was not only leading several major business deals but was also responsible for the nurturing of Fujitsu's data scientists. In 2014, I returned to Fujitsu Laboratories where I started full-scale research on AI. In 2020, I became a fellow of Fujitsu. Besides my activities at Fujitsu, I am also taking part in national committees as an AI specialist and am also active as a visiting professor at several universities.

6. Can you mention some major achievements of Fujitsu under your leadership.

Currently, our research on AI is conducted on a global scale under a global management system. We are successfully cooperating with many players worldwide and are conducting R&D with members globally from offices in places including the USA, China, UK, Spain, Canada, France, Israel, and Japan. We are also actively promoting industry-academia collaboration, through working with internationally renowned universities. We think it is very important to go forward in AI research through open innovation and are striving to compete globally in R&D in AI. In addition to technology, there are different national regulations and cultural differences pertaining to AI ethics. Our aim is to absorb and deploy them in our research of an AI that can be applied on a global scale.

Personally, the base of my research at Fujitsu has always been AI, and I feel that having been able to conduct AI-related R&D for so many years is one of my greatest personal assets. My goal has always been to contribute to the creation of new value and the creation of reliable, ethical AI that can be applied on a global scale. By combining research in new technology areas like quantum computing, I hope to continue contributing to the creation of technologies unique to Fujitsu that play a role in solving some of the challenges facing the world today.

I am proud of the progress we have already made in addressing some of the challenges surrounding issues like AI quality and AI ethics. The "Fujitsu Group AI Commitment," has been an important step toward a non-biased, ethical AI. We are also leveraging our AI technologies to make peoples' lifesavers, as demonstrated by our tsunami prediction technology. Our AI technology also supports more advanced decision-making of humans, as demonstrated by our cancer treatment technology in the healthcare sector for example.

My personal goal is to further strengthen our AI R&D organization to realize an ethical, trustworthy, and secure AI that contributes to the safety and wellbeing of even more people and their lives.