NOTE: this is an archived page and the content is likely to be out of date.
Vol. 54, No. 5, October 2018
Fujitsu Laboratories celebrates its 50th anniversary in November 2018, which was established in 1968 as an organization independent from Fujitsu Ltd. for conducting the world’s top-level technological development in a free atmosphere for researchers. Since then, Fujitsu Laboratories has been leading the world in research and development by producing a succession of innovative results and providing new value to people, society, and businesses. This issue presents the latest R&D results in eight emerging technologies with which Fujitsu Laboratories aims to lead the world, and introduces current activities in Digital Co-creation.
As a technology company, Fujitsu always tries to be a powerful leader of innovation with the aim of providing a prosperous and dream-inspiring future to the people in the world through ICT. As a member of Fujitsu Group, Fujitsu Laboratories has a mission to put Fujitsu’s corporate philosophy into practice and continuously drive its growth through cutting-edge technologies. Today, digital technologies are becoming indispensable not just for social innovation but for corporate development, and a new growth model for digital transformation is required. Therefore, we have to develop cutting-edge technologies and associated business models as well as to fuse high-level expertise for dealing with digital technologies with a wide range of skills for utilizing expertise in business activities of Fujitsu customers. By providing technologies and business models by top-level experts, Fujitsu Laboratories is promoting Digital Co-creation to create new values in collaboration with customers. This paper describes Fujitsu Laboratories’ R&D strategies for Digital Co-creation, medium- and long-term approaches in cutting-edge technology development, and recent research achievements.
Since the development of practical stored-program computers in the late 1940s, performance has risen amazingly by about 1012 times over a period of 70 years. However, it is generally recognized that semiconductor transistor scaling is reaching its limits and that Moore’s law is coming to an end. Regardless of these technical issues, the explosive increase in the amount of data generated in today’s IoT era is expected to continue, and it is highly anticipated that this data will be used to create new value and novel services. Meeting these expectations will therefore require improvements in performance independent of Moore’s law. To address these issues, Fujitsu Laboratories proposes domain-specific computing as a new computing paradigm. The aim of domain-specific computing is to break through Moore’s law by adopting architecture specific to the type of processing needed in fields such as knowledge processing whose objective is not to obtain rigorous numerical results. For example, in application to deep learning engines, high-speed image search engines, and machines dedicated to combinatorial optimization problems, domain-specific computing has demonstrated that it showed 50–12,000 times higher performance than that of conventional approaches. In this paper, we describe the direction of domain-specific computing as a new computing paradigm and present specific application examples.
AI technology is playing a key role in the digitalization and transformation of many aspects of business and society. Since people support large segments of business and society, this transformation needs to be driven by an optimal combination of the best features of people and AI and the provision of high value. Fujitsu Laboratories is researching and developing AI technology on the basis of a human-centric way of thinking. The objective here is not just technology that can demonstrate a level of performance equal to or greater than experts but also that to explain AI output results, which is essential for establishing a close collaboration between people and AI. This paper describes an overall view of FUJITSU Human Centric AI Zinrai. It then introduces advanced technologies that focus on learning such as Deep Tensor, knowledge processing such as knowledge graphs, and “Explainable AI” that combines those technologies using practical examples in fields such as medical care and security.
Data distribution and utilization are being consolidated due to support by legislation such as the Basic Act on the Advancement of Public and Private Sector Data Utilization and the Amended Act on the Protection of Personal Information and also by the establishment of the Data Trading Alliance consisting of industrial, governmental, and academic institutions. As corporate businesses become increasingly digitalized and connectable with potential partners in the future, Digital Co-creation among different industries will accelerate more than ever and give rise to innovations. Fujitsu Laboratories propose a concept, Connected Digital Place, which enables Digital Co-creation among various industries. As a core part of the concept, we have been conducting research and development of a data-driven platform that manages data, converting its formats into connectable ones. This paper introduces the latest research to realize the data-driven platform, and Fujitsu’s approach to promoting the value creation cycle through the use of this platform.
As the 5th Generation Mobile Communication System (5G) era approaches, various efforts are underway to realize cyber-physical systems (CPS) in which real world people/things and the cyber world come together. Fujitsu Laboratories conducts R&D on key technologies to facilitate the social implementation of diverse CPS services and enable experiences never before possible. The “field area management platform” makes possible control of various devices in the real world without having to develop separate technology for each device. The “real-world service platform” enables the efficient implementation, in line with real world states, of the combining and recombining of services that handle large amounts of data by mapping the real world as “virtual objects” into cyberspace. The “environmentally adaptive virtual networking technology” makes possible the construction and control of secure network environments based on the attributes of users and their purposes. This paper describes the two above-mentioned platform technologies, which are important for connecting the cyber and the physical worlds, and the networking technology to connect the real world and services in a timely and secure manner.
All kinds of things are becoming digitalized, and the world of Digital Co-creation that combines these things to create new businesses is expanding. For companies to survive and grow their business in this world, they must create value by presenting their strengths to other companies in the form of application programming interfaces (APIs) and at the same time make use of other companies’ strengths by way of APIs. This requires engineering technology for quick and reliable system development/operation centered on APIs. Fujitsu Laboratories calls the technical field in which this is achieved API-Publish & Compose. The Laboratories engages in research and development of system analysis technology for the purpose of making APIs of a company’s particular strengths extracted from existing complex systems. We are also working on development and operation technology for quick and safe systematization of businesses using APIs offered by other companies. This paper describes the activities of Fujitsu Laboratories with regard to technologies required by both API providers and API users, and application examples.
As the value brought about by data utilization is attracting attention, the global big data market has been making growth at an annual rate of more than 10% and is estimated to reach a scale of about 20 trillion yen by 2020. In line with this trend, development of laws for promoting proper data distribution and utilization has progressed on a global basis, and technologies to meet the requirements of those laws are beginning to come into practical use. However, there are also anxieties expressed arising from the inability to make decisions regarding risks in personal data distribution; individuals may agree to data distribution without realizing how high the risk is, or business owners may cause privacy issues by distributing personal data with low anonymity, possibly resulting in major losses such as compensation for damages. To deal with these issues, Fujitsu Laboratories has developed a technology to quantify privacy risks from personal data disclosures in terms of monetary value. We have also developed a model for calculating the identifiability (how low anonymity is) of data after anonymization, which was insufficient in the past, and confirmed that these are applicable to real data. Furthermore, we have developed a high-speed identifiability calculation technology that allows for the calculation of data sets on a scale of 1 million people in about an hour with a general performance PC, confirming adequate practicability. This paper describes the technology that allows for risk evaluations regarding personal data and the concept of realizing a society that can better extract the value of data by utilizing this technology.
The Japanese working-age population is estimated to fall to about half of the total population by 2060. Along with this change in society, new working lifestyles such as working remotely and co-working with robots are becoming popular. Coming together with others within the workplace in the real world can create advantages for organizing our work life, but current computer technologies—including robots and other autonomous artificial intelligent systems—have not helped to create the same advantages in remote working spaces yet. Sensecomputing Research Center in Fujitsu Laboratories takes on this research issue regarding people’s need for relatedness. We aim to develop computer-supported cooperative remote working spaces that satisfy people’s basic psychological need for relatedness. In this paper, we introduce a preliminary study on building design for future remote working spaces. This paper presents findings obtained on how a user interacts with other persons using his/her interpersonal relationship building behaviors in virtual reality settings. It then shows that future remote working spaces will let a user be informed that another person would like to approach or avoid him/her, using where the other person’s body parts are positioned, how angled the other person’s upper body is, etc. This paper also includes the background for the study and the preliminary results of user observations, including its methodology and system configurations.
New materials, so far, have generally been developed through repetition of numerous experiments based on researchers’ experience and knowledge. Recently, materials informatics (MI), which allows for the identification of guidelines for material design on computers, is attracting attention along with improvements in computer processing capacity and advances in information science and technology. Fujitsu Laboratories has been conducting research on this MI technology with the aim of creating new materials and devices and new drugs. MI is based in data, and the challenge is how to exhaustively gather and utilize good quality data. Fujitsu Laboratories is taking the approach of making the enormous amounts of accumulated material and device data into databases for analysis with AI technology. Meanwhile, we are also studying informatics-based design using data and AI by applying simulation technology, at which we excel, to generate the data on computers with a degree of accuracy equivalent to that of experiments. This paper presents the realization of an IT-based drug discovery that makes use of molecular simulations based on physics and chemistry, as well as the optimum design of magnetic devices that combines magnetic simulations and AI technology. It then presents informatics-based design technology aimed at new material development. It also describes the MI approach utilizing experiments, analyses, and simulations as an MI technique that includes the process as well as material search and presents future prospects.