The Fujitsu IoT and Big Data Forum was held on November 5, 2015 at Bellesalle Shiodome in Tokyo. This report describes how the Forum showcased the pioneering work of the Fujitsu Group in Big Data and IoT as well as the increasingly topical field of artificial intelligence (AI).
IoT and Big Data Drive Innovation
The event opened with a presentation from Executive Vice President Hiroyuki Sakai on the Fujitsu vision of the digital revolution.
He argued that the digital innovation already constitutes a key issue throughout industry. He cited the example of a joint project between Fujitsu and Omron to integrate multiple sources of production line data into user-friendly readouts, which resulted in the client being able to locate bottlenecks around six times faster than before. Another example was a Smart Agriculture solution developed for the City of Iwata that integrates expertise in farming and food production from a variety of different sources, from seedling suppliers to local government agencies and academic institutions.
Mr. Sakai also used video clips to describe other major Fujitsu programs, such as the MetaArc cloud-based digital business platform, which uses the latest technology to help clients harness the benefits of the digital innovation; Zinrai, the Human Centric AI system that represents the culmination of over 30 years of research into AI; and open innovation achievements. Fujitsu is committed to working with clients in the pursuit of innovation in the digital business space.
The next keynote speech was delivered by Project Professor Shuichi Inada from the Research Center for Advanced Science and Technology (RCAST) at the University of Tokyo. Quoting a recent report that claims that about half the jobs in the U.S. will be automated within the next couple of decades, he argued that computers designed for high-capacity data analysis are now superior to humans at performing a range of tasks requiring intelligence. This means that our jobs will shift towards more deeper intellectual tasks that are still beyond the reach of machines, such as design and process optimization in manufacturing. According to him, the key to achieving true innovation in the age of the Internet of Things and data-driven innovation is to think beyond mere data collection, to realize that the ultimate goal is creating experiences or conveying emotions.
"Data is not enough in and of itself to achieve innovation," Professor Inada said. "Data certainly provides the foundation for innovation, but you cannot predict the outcomes in advance. Innovation is not a cost-benefit process. The results become apparent as you proceed, and you edge closer to your goal. The business world is changing and restructuring rapidly, and we cannot expect to be able to keep up if we operate in isolation. We must pool our wisdom and technical expertise to implement reform and change that are in keeping with the times. This is the way to do business in the 21st century."
AI, Cars and Transportation Systems: Leading Examples of How Big Data and the Internet of Things Are Transforming Our Lives
The keynote speech was followed by a seminar session on recent leading examples of how Big Data and the Internet of Things are being used in fields as diverse as manufacturing, marketing, medical care, cars, transport and public infrastructure.
Using the Internet of Things to create safer workplaces and public spaces through innovation
The role of sensor data analysis in workplace processes and monitoring
Hiroyuki Goto, Director of the IoT Business Promotion Division, described the vision on the human-centric Internet of Things where Fujitsu works alongside clients to promote human safety and security. He introduced UBIQUITOUSWARE as an Internet of Things package for analyzing sensor data from a wide range of sources including location information and so-called vital data and extracting data of genuine value. UBIQUITOUSWARE uses unique algorithms to form a three-dimensional picture of the wearer in terms of parameters such as orientation, direction of movement and walking speed.
Mr. Goto cited two examples of UBIQUITOUSWARE: the Vehicle ICT FEELythm for cars, a sensor that clips on the driver's ear and monitors pulse waves for signs of drowsiness; and a workplace safety solution called a "Vital-Sign Sensing Band" that monitors pulse rates and physical movement and provides alerts to prevent possible heatstroke.
The future of artificial intelligence (AI)
Fujitsu has been conducting research into artificial intelligence since the 1980s, culminating in the recently released Zinrai. Zinrai features perception, recognition, knowledge formation and decision-making technology along with highly advanced learning and research functionality. Fujitsu Laboratories Director Yuki Hara explained that AI is often associated with concerns about machines taking over and controlling humans. However Fujitsu espouses a human-centric approach to AI, where the central focus is on the human operators. Fujitsu is also focused on the idea of AI that becomes more intelligent through growth. According to Mr. Hara, AI can be incorporated into solutions and services in an ever-expanding range of fields such as manufacturing, urban transportation, medical care and finance.
To assist corporate clients in identifying the optimum AI solutions for their needs, in December 2015 Fujitsu opened a new AI Consulting Division where AI technical consultants, backed by some 200 AI researchers, engineers and curators, will work with clients to develop optimized AI scenarios.
Next to address the Forum was Seishi Okamoto, Project Director of the Knowledge Information Processing Laboratory at Fujitsu Laboratories, who delivered a presentation on Zinrai, a real-life application of AI technology. Zinrai features three core technologies: Mood Media, which replicates the five sensory inputs of humans in order to understand and interpret human emotion; Knowledge, which generates the knowledge that forms the basis of both human understanding and machine processing; and Mathematics, which uses supercomputer technology for mathematical modeling of key challenges in industry and society. Zinrai also features a continuous learning system that enables it to grow and develop by identifying useful knowledge and patterns within everyday interactions.
There was considerable interest in the presentation on Deep Learning, the latest manifestation of neural network (NN) technology, a form of mathematical modeling that seeks to replicate the nerve nets of neurons and connecting synapses in the human brain. Project Director of Fujitsu Laboratories Knowledge Information Processing Laboratory Seishi Okamoto described how Fujitsu Laboratories has developed a unique NN model comprising 1.1 million neurons and 730 million synapses. He gave a demonstration showing how the model automatically acquires the characteristics of object recognition as a means of identifying objects within images to a remarkable degree of accuracy. Mr. Okamoto also presented several applications of AI technology, including Mood Media technology, which can identify emotions or intentions in speech and recognize signs of stress; high-resolution supercomputer simulation of tsunami inundation depths with high-speed parallel mathematical computations; and the EvaCva local information service that combines linked open data (LOD) with machine intelligence.
Smart mobility in the growing world of the Internet of Things
Innovation Business Unit Head Hirohisa Yamaguchi talked about applications of the SPATIOWL mobility platform in areas such as traffic safety and security and urban planning and design, as well as the progress of recent validation trials. SPATIOWL is a Fujitsu cloud service that allows location information from sensors, vehicles and individuals to be aggregated in a central ICT infrastructure database, typically in the cloud, for the purpose of analysis. By combining these different types of data it is now possible to create value-added services. SPATIOWL is being used in Japan and overseas in the development of safe and reliable urban planning and transport solutions.
He cited four applications of the SPATIOWL platform: a monitoring service for families with elderly relatives; an open data residents' service; a pedestrian flow and behavior analysis system designed for marketing and business optimization; and an urban mobility and traffic management solution. The City of Kawasaki has developed a service for parents as part of a drive to make Kawasaki the most family-friendly municipality in Japan. Residents had complained that information about parenting and family events was difficult to locate, with events spread across multiple sections and departments, so the municipal authorities developed an app specifically for parents based on SPATIOWL. Public-domain information about municipal facilities and family-oriented events is formatted to enable intelligent machine analysis. The system identifies information about events and destinations relevant to parents, which is then combined with SPATIOWL position information and made available via the smartphone app. The app tailors all information to the child's age and residential address to ensure that only relevant local information is presented.
Overseas applications of SPATIOWL include a traffic monitoring smartphone app in Indonesia, a bus probe data system in Taiwan and a system that identifies traffic congestion from video images in China. In Singapore, meanwhile, SPATIOWL is being used to alleviate traffic congestion in trials of an ICT system that provides real-time aggregation of Big Data from multiple data sources including traffic monitoring, human behavior and open data. The system displays a constantly updated list of suggested actions and forecasts that are ultimately designed to reduce traffic congestion by encouraging drivers to make different choices on the road.
Mr. Yamaguchi concluded his presentation with a statement about combining client services and ideas with Fujitsu technology to create business ventures through collaboration.
The exhibition space featured six demonstration areas showcasing the latest IoT solutions from Fujitsu, such as UBIQUITOUSWARE based on sensing technology and the SPATIOWL real-time position information system for people and transit systems along with models of IoT applications. There was also a section devoted to the Business Application Operational Data Management & Analytics solution, featuring a Big Data analysis engine equipped with Zinrai machine learning technology, which illustrates how the Internet of Things will soon be an integral part of everyday tasks and processes. The demonstration areas proved popular amongst visitors keen to gain a first-hand experience of the latest IoT solutions and technologies.
UBIQUITOUSWARE IoT solutions
Three UBIQUITOUSWARE devices were presented. The Vital-Sign Sensing Band is a safety solution that helps to reduce the incidence of workplace accidents and incidents in the construction, manufacturing and agriculture industries through continuous monitoring of worker health status, orientation and movement. Location badges and tags provide accurate tracking of pedestrian traffic and location information for inanimate objects in hospitals and commercial facilities. And the remote monitoring station can be used to monitor the condition of an elderly relative based on everyday sounds such as speech, coughing and breathing patterns during sleep.
Intelligent production line analysis
A factory visualization system has been developed based on trials conducted at the Fujitsu I-Network Systems Yamanashi plant. Monitors indicate the status of printed circuit board production lines through a series of digitized graphs that can be used by administrators, factory supervisors and workers to identify and evaluate production bottlenecks in real time.
What is applied to this initiative is the IoT solution with the theme of "factory visualization." Correlations between sensor data from the production line and equipment logs, production schedules, production figures and operator observations are aggregated in the cloud and used to provide a visual representation of the overall status of the facility, while analysis and prediction techniques are used as the basis for timely modifications and improvements to production processes.
Predicting faults through anomaly detection
The Big Data solution known as Operational Data Management & Analytics, released in April 2014, features machine learning, a core element of Zinrai. Machine learning significantly improves the accuracy of fault prediction in equipment and services by using a technique known as anomaly detection. An anomaly refers to a state or condition that differs from the standard state or condition. Once the machine learning has been taught the normal states or conditions, it can flag differences as being faults, providing the basis for change forecasting. In this way, the solution contributes to reforming client processes and creating new business, such as equipment maintenance at factories and production plants, automation of equipment maintenance, and operational continuity improvements.
Omni-channel digital platform
The Internet of Things can be used to encourage consumers to buy by sending notifications of discounts or limited offers at a particular store to smartphones in the vicinity. In this model, location information for the device is detected via wireless LAN, and the customer's pre-registered attributes are used to determine whether the offer details match the personal preferences of the customer and provide personally tailored notifications. Notifications are pushed to the device in real time through the dedicated app. This demonstration showed how the distribution industry is increasingly using digital technology as a marketing tool to gain better access to potential buyers.
Real-time location information for people and vehicles
The demonstration showed a real-life application of a SPATIOWL smart mobility solution for regulating pedestrian traffic flows at major events. An on-screen display of traffic conditions near the venue directly after the conclusion of the event is used to show people in the vicinity their best transport options. It is also used by the control center to send security personnel real-time updates of pedestrian traffic flows and transport conditions, and to determine the optimum deployment of security personnel in and around the venue.
Using the Internet of Things for factory optimization
The Factory Performance Dashboard displays a comprehensive overview of production data that provide an accurate snapshot of the current status of factory operations. Data types that are normally handled separately, such as power consumption, production output, quality standards, inspection records and equipment maintenance records, and environmental and management data such as unit cost calculations, are aggregated with sensor data and presented on the dashboard in a federated format that provides a comprehensive overview of factory operations. The presenter showed how the dashboard can be used to identify faults and areas of concern.
Virtually all the presentations and demonstrations at the Fujitsu IoT and Big Data Forum were packed to capacity, in a potent reminder of the level of interest in Big Data and the Internet of Things. Many people headed straight to the demonstration area after listening to a seminar, eager to try out the technology first-hand. The presenters were flooded with questions, indicating a high level of interest and engagement all round.