News and Insights

“Robots and humans can make a good team.”

When it comes to solving highly structured problems, robots are often already better than humans. To take the example of agriculture, robots can continuously monitor parcels of land to ensure the optimal use of water, fertilizers, or pesticides.

Many businesses that have mastered these and similar technologies provide measurable added value and therefore make attractive investment propositions. However, robots are not up to tackling complex workflows or interacting with humans just yet.

Professor Siegwart, how has Switzerland managed to develop into a global melting pot in the field of robotics?

Roland Siegwart: Robotics is a systems technology that combines precision engineering, sensors, actuators, and intelligence in a complex machine. The Swiss economy and Swiss research have a long tradition in all these areas – a tradition on which the universities’ research laboratories have been able to build. By recruiting professors and staff in robotics at ETH Zurich and École polytechnique fédérale de Lausanne (EPFL), and thanks to various flagship programs such as NCCR Robotics, NCCR Digital Fabrication and Wyss Zurich, a melting pot for outstanding robotics research, technology transfer, and start-ups has developed that is unmatched anywhere in the world.

Competition keeps business on its toes – does this also apply to the universities involved in robotics research?

Roland Siegwart: Academics are ambitious and love competition. Like top sportsmen and women, they always want to take first place and be the best. This boosts the ability to innovate. That said, they also enjoy close working relationships with each other. In ETH and EPFL, Switzerland not only has two of the leading universities in the world, but also two institutions with some of the best international networks.

Which universities or researchers do you work with most closely?

Roland Siegwart: I’m in very close contact with all the main robotics researchers, and more than 20 nationalities are represented in my laboratory. We work especially closely with MIT, the University of Sydney, Caltech, and the CSIRO in Brisbane, for example. As part of European projects, we cooperate with the University of Freiburg, the University of Naples, RWTH Aachen University, the EPFL, and many more. We are involved in various research projects with companies such as ABB, Microsoft, Huawei, and Intel.

Our partnership with the ETH gives us access to research projects or spin-offs in need of funding. Investments in spin-offs can be of particular interest to our Asset Management arm.

Filippo Rima

Credit Suisse Asset Management has endowed a chair in robotics at ETH Zurich. How important are partnerships like this?

Roland Siegwart: These partnerships are really important. The support of Credit Suisse enables us to build on ETH’s position as one of the world’s leading universities in the field of robotics.

Filippo Rima: More than 160 years ago, ETH and what was then Schweizerische Kreditanstalt (now Credit Suisse) were established at more or less the same time to build the Gotthard tunnel. In the same vein, we also need partnerships today so that we can press ahead with forward-looking technological developments such as robotics.

What are the benefits of the partnership for you, Mr. Rima?

Filippo Rima: It creates an ideal basis for the ongoing exchange of expertise and gives us access to research projects or spin-offs in need of funding. Investments in spin-offs can be of particular interest to our Asset Management arm. I’m thinking of equity funds that invest in specific themes such as global automation here, an area in which robotics plays a key role.

We are proud to be helping the university to build on its leading position. What are the benefits of the partnership for you, Mr. Rima?

Filippo Rima: It creates an ideal basis for the ongoing exchange of expertise and gives us access to research projects or spin-offs in need of funding. Investments in spin-offs can be of particular interest to our Asset Management arm. I’m thinking of equity funds that invest in specific themes such as global automation here, an area in which robotics plays a key role.

We might say that robots have grown up within industry. How is the industrial robots segment doing, Mr. Rima?

Filippo Rima: According to data from the International Federation of Robotics, more than 380,000 industrial robots were sold in 2017. The market also looks set to see double-digit growth over the next few years, contributing to the competitiveness of whole sectors of the economy. However, industrial robots are reaching their technological limits. The industrial robots of today can’t cope once production workflows become less structured and products are switched quickly. They’re unable to learn new process steps themselves or autonomously adapt to new situations.

What are the implications of this, Professor Siegwart?

Roland Siegwart: Robots need to become more flexible and “intelligent” for them to be able to adapt to new circumstances. They are increasingly being equipped with cameras and other sensors to enable them to analyze situations and react accordingly. As a result, robots can also be used to automate small batch sizes and take on laborious tasks otherwise performed by humans. But this step is highly complex, and many challenges still need to be overcome in this area.

Are robots that perform services still in their infancy?

Roland Siegwart: That’s one way of putting it, yes. We all know it’s robots that build our cars, but have you ever seen a robot that can repair your vehicle?

For the next generation of industrial robots – and also for service robots – we need new technologies that permit all-round perception and analysis of the environment alongside tactile interaction. Tasks that often appear simple to human beings like clearing the table after a meal are still inconceivable for robots today, and will remain so in the next few years. On the other hand, robots working on production lines can place parts with submillimeter accuracy, something which is beyond us humans without the necessary assistance. So the future belongs to collaborative robots. Human beings will take care of more interesting work for which the focus is on understanding, creativity, tactile ability, and interactivity, while robots will carry out repetitive tasks that require precision and the ability to do things over and over again.

Anymal the model canine

Four-legged robots like the ones created by the young ETH spin-off company Anybotics are clearly superior to wheeled models, as they can navigate rough terrain and climb stairs. The robotic dog, which weighs about 30kg, can be used to take measurements on oil rigs, conduct land surveys, and in rescue operations. Anymal can also act as an obedient companion on walks and small excursions.
anybotics.com

What skills will service robots learn over the next few years?

Roland Siegwart: The first requirement for service robots is reliable and robust navigation. This means that robots need to be able to independently draw up maps of their environment, locate themselves within it and move around purposefully but collision-free. A great deal of progress has been made in this area in the last few years. Robots can now create mapsusing cameras and lasers and then move relatively reliably. The key thing now is to make these technologies fit for industrial and dayto-day use over the coming three to five years so that they can be deployed in driverless vehicles, cleaning robots, or drones.

Mr. Rima, why are robots of interest to investors?

Filippo Rima: As the cost of technologies used in automated systems is falling continuously, robots are increasingly being used in areas of everyday life.

Robots are being encountered more and more often in shops, restaurants, and offices. Not only that, automation has now entered the realm of hospitals and government offices as well as cars, trains, and aircraft, not to mention our homes. From the investor’s perspective, the increased use of robots in industry and the growth in automation in many other areas of the economy are creating long-term investment opportunities. In a world in which global growth is broadly in decline, investors are becoming more and more interested in areas exhibiting structural growth.

As an investor, how can I identify the right companies?

Filippo Rima: It’s not easy to identify businesses with the highest growth potential, especially as this market is developing very dynamically. Because many specialist companies are not listed and can be unforthcoming in providing information, it can be very hard for “normal” investors to find the best opportunities. On top of this, products are becoming ever more complex, which means that considerable expertise and experience are required to evaluate them. Funds managed by specialists are at a clear advantage over individual investors in this area.

What factors are limiting the rapid spread of deep learning robots?

Roland Siegwart: “Deep learning” usually refers to learning algorithms that roughly replicate our current understanding of the brain, though obviously to a very limited extent. Non-linear functions such as the region represented by a street or a field are learned using a neural network and a great many training examples. Significant progress has been made in recent years in relation to one-dimensional problems such as analyzing medical image data. Computers can now identify tumors in images more reliably than their human counterparts.

In the near future, robots will help to make agriculture far more sustainable as resources like water or fertilizers will be used to best effect and pesticides will be administered in precise doses.

Professor Roland Siegwart

And how are things looking in the area of multidimensional problems?

Roland Siegwart: Multidimensional learning of complex interrelationships requires millions of training examples and greater computing power to the order of several dimensions. Deep learning as it is understood today is not capable of this. In its current form it still needs a defined objective. It’s very difficult to provide this type of definition for complex workflows.

Presentday deep learning algorithms are not yet much more than programs that enable large data streams to be optimized and analyzed. For example, deep learning makes it possible to identify cancerous tumors (output) based on images (input). Computers are better at this than people as they can access and process large volumes of data much more quickly. But the abilities of artificial intelligence (AI) are still very limited as things stand. It’s therefore a very bold claim to extrapolate AI systems that solve structured and narrowly defined problems to robotic systems expected to tackle the highly complex, multimodal problems we encounter in our everyday lives.

If we fail to make great progress in agriculture and the logistics of distribution, large parts of the world’s population will continue to suffer undernutrition and malnutrition. How can robots contribute to resolving this problem?

Roland Siegwart: There’s great potential for robots to be deployed in agriculture. Robots can continuously monitor fields and intervene immediately if, for example, more water or fertilizer is required or pests need to be removed. In the near future, this will help to make agriculture far more sustainable as resources like water or fertilizers will be used to best effect and pesticides will be administered in precise doses. We expect a fraction of the pesticide volumes used today to achieve the same effect, and it should be possible to do much of the work involved in combating pests “mechanically.” At present, around 30% of food is lost before it even leaves the field, while another 30% is lost during distribution and storage.

It has been proved that robots can learn from human beings. But can humans learn from robots?

Roland Siegwart: There are not many lessons people can learn from robots in their everyday lives just now. But people can aim to develop an optimal working relationship with them, as robot and human skills can complement each other. Robots don’t get tired, they can carry out highly precise movements, and they can carry heavy loads. Human beings are unbeatable when it comes to analyzing complex systems, interacting with other people, and generating new ideas.

Filippo Rima (laughs): I’ve come to the conclusion that we can certainly learn something from robots. Discipline, hard work, precision, and the ability to work under pressure are all virtues that we as people could do with a little more of.