Kathy Gibson is at Fujitsu Forum in Munich – As Fujitsu builds a new business model and strategic directions, Fujitsu Laboratories has shifted its focus to bring new technologies in quantum computing and artificial intelligence (AI) to market.
The company aims to follow a new business model as it switches to being a services and solutions company. “This means Fujitsu will change from being an IT company, to a digital transformation company,” says Hirotaka Hara, CEO of Fujitsu Laboratories.
He says the company aims to take a path to achieving this goal that takes advantage of its strengths – including technology and business know-how, combined with new imperatives.
Fujitsu Labs is working on developing solutions in computing, AI, 5G, IoT, cybersecurity, cloud and data. And these are informed by the pillars of creating trust, leading with digital and acting globally, Hara says.
“Today, product reliability alone is not enough,” says Hara. “People also require trust as a new element.”
Digital trust has to be built into the technology and also needs to impact on the physical world with safety, reliability, explainable AI, eligibility and immediacy.
“Trust is a key element of penetrating AI in the real world,” says Hara. “Fujitsu was one of the first to participate in discussions of AI ethics and is one of the founding members of the Framework of European Commissions AI ethics guidelines.
The five ethical principles of AI are:
* Provide value to customers and society with AI;
* Strive for human-centric AI;
* Strive for a sustainable society with AI;
* Strive for AI that respects and supports people’s decision-making; and
* As corporate social responsibility, emphasise transparency and accountability for AI.
AI will be accomplished via superior technology, Hara adds. Fujitsu is working on new computing technologies including digital annealer and content-aware computing.
“This new direction, content-aware computing does not depend on hardware,” he explains.
Content-aware computing technology works by analysing the content of computation to automate trial and error.
The new systems will be well-suited for deep learning, where they will perform up to 10-times faster than existing systems.
AI challenges that Fujitsu Labs is working on now are explainable AI (XAI) and AI quality and it will be launching solutions in these areas in the near future.
The company also partners with some of the world’s leading research institutions to bring AI solutions to market.
One of these is Inria in France, where the company is working on AI and topological data analysis (TDA) for time-series data.
The organisations will soon, jointly announce two solutions from its collaboration efforts and will make them available via open source.
Dr Jean-Frédéric Gerbeau, deputy CEO for science at Inria, explains TDAs are mathematical models and algorithms used to extract topological features of data.
Fujitsu and Inria are working on joint research to develop innovative technologies used on TDA for complex data – with a particular focus on complex time-series datasets.
Dr Gerbeau says time series data is about measures, depending on the time. Use cases can be found in wearables, biological applications, sound/vibration and finance.
The benefits of TDA is that it is robust in regards to individual differences and well adjusted to be less dependent on individual variability.
The solutions developed with Fujitsu Labs have been used for damage detection on bridges, analysis for aircraft blades and the early detection of arrythmia.