People

Sony AI

Attracting the World's Best AI Researchers and Engineers to Unleash the Potential of Sony's AI

Aug 31, 2020

Sony AI has been established as a wholly owned subsidiary of Sony Group in April 2020. The organization has offices in the U.S., Europe and Japan, aiming to collaborate with relevant Sony group businesses, to boost Sony's AI technology portfolio, in a unique way. Michael Spranger, COO of Sony AI and Peter DÜrr, Director at Sony AI ZÜrich, each talked about their mission, thoughts, and their R&D activities at Sony AI.

Profile

  • Michael Spranger

    COO
    Sony AI

  • Peter DÜrr

    Director
    Sony AI ZÜrich

Promoting and Co-working on Three Flagship Projects
within Sony Group

──What is Sony AI's objective of establishment and the company's mission?

Michael Spranger:Sony AI is committed to achieving its mission to "Unleash human imagination and creativity with Artificial Intelligence." In particular, we think that AI will help to empower the imagination and creativity of artists and creators around the world. This mission is aligned with Sony's Purpose: "Fill the world with emotion, through the power of creativity and technology."

In collaboration with Sony's business units and R&D Center, we are currently working on three flagship projects that are set up in a way that combines the deep technology stacks from specific businesses and R&D units, and Sony AI will add specific AI technology capabilities. The three projects take place in the field of Gaming, Imaging & Sensing, and Gastronomy. Gaming and Imaging & Sensing are promising growth drivers for Sony's businesses and Gastronomy is a completely new domain for Sony that offers great potential for technical explorations. In the meantime, we are also actively working on adding flagship projects, to further extend the assets Sony already possess and contribute to the future success.

──Why is gastronomy so important to Sony's AI initiatives? How is it connected to AI?

Spranger:We believe gastronomy and AI have a great potential to become a new area of activity for Sony. In fact, Sony has a history in gastronomy, as its founder, Masaru Ibuka's factory produced rice cookers at some point before the establishment of the company. Also in the 1960s, the Sony Ginza building featured one of the first authentic French restaurants in Tokyo - a branch of Maxim's de Paris.

Today, we are trying to bring AI and robotics to gastronomy. This project is composed of two parts. On the one hand, we are working on AI technology that allows chefs and artisans to create a new, more healthy, sustainable and tasty dishes. On the other hand, we are working on robots that can become new tools to prepare delicious dishes in the restaurant or at home. All of these activities benefit from Sony's amazing portfolio of technology in sensing, robotics but also in creative AI.

──What strategies does Sony AI have for being active and Sony-unique in the front lines of AI research and application?

Spranger:Research and development of AI is a fast-growing field, and the competition has become more intense than ever. A question that everyone is asking is: how can we create a competitive advantage? The answer is simple. Hire the best talent in the world and allow them to work on exciting, ground-breaking projects. The flagship projects that Sony AI can carry out by leveraging Sony's assets in gaming and imaging and sensing — as well as in music, movies, and electronics— is what will really distinguish us from competitors in the global AI research and development landscape. Sony is the only place where AI researchers can focus on their research, while being able to work very close with creators in such an interdisciplinary field.

AI is a ubiquitous technology that will affect across all areas of Sony business from financial service to sensing to entertainment. As seen from the recent announcement of the Intelligent Vision Sensor with AI processing functionality, AI presents an opportunity for Sony to transition from being a sensor producer to an AI integrated sensing solution provider that could potentially support the next generation of cars, manufacturing robots, and kitchen appliances, to name a few. AI will play a transformative role when creating content such as movie, music, and games allowing more people to express themselves through technology. We will see new shapes of creativity emerge and whole new fields of music, games and also gastronomy -- driven by AI technology.

To stimulate creativity in Sony and allow businesses within Sony Group to push the state-of-the-art technology quickly, Sony AI will also actively seek collaboration with external partners. To develop and use this technology is not just Sony AI's mission; it is an important goal for Sony group as a whole. So, my hope is that together, we can make Sony the best AI technology company in the world.

The dawn of the next generation AI

──What are you working on at Sony AI ZÜrich?

Peter DÜrr:At Sony AI ZÜrich, we are currently working on combining new sensor technologies developed at Sony Semiconductor Solutions with new machine learning methods and robotic actuators. If we can unlock the full potential of these sensors by developing new learning and control algorithms, as well as the computing hardware to handle the sensor data properly, I believe that the resulting systems will open up new applications in areas such as factory automation, autonomous driving, or drones.

──What is your insight on the way future R&D of AI ought to be?

DÜrr:Our perception of AI today is very much influenced by deep learning-based classifiers, which use training data that is labelled in advance. The next generation of AI systems, however, will not rely on such carefully prepared data. They will generate the necessary data by computation, for example by using simulations, or gather it by interacting with the environment, which is the domain of robotics. Therefore, we should not treat AI and robotics as entirely separate research topics.
Moreover, in the very near future, the learning algorithms for AI will no longer be constrained to big servers. Computation will happen very close to the sensors, which will enable new applications through lower latency, lower power consumption, and enhanced privacy.

Related article