Bruno Zamborlin, HyperSurfaces - on how students make their startup

Bruno Zamborlin - Head and founder of startups HyperSurfaces and Mogees, UK.

HyperSurfaces - startup,

developing user interface technology based on the recognition of changes in vibrations (vibrations) of any hard surface.

Mogees - A startup that develops the technology of using any surface as a game or music interface.

Make any surface an interface

- The innovation of HyperSurfaces is to recognize the vibration of a hard surface. Tell us how it works.

- Vibration is a phenomenon that occurseverywhere. Every time people interact with a physical object, touch it or move it, the object emits a signal. It is unique for each movement, so using the simplest vibration sensor, you can catch it, and then using the algorithm, artificial intelligence and machine learning in real time to recognize what is happening with the object. The idea behind HyperSurfaces is to recognize in real time how people interact with objects around them.

- That is, any object - a wall, a machine, a table or a hanger - can work with this technology? What other surfaces are available for work?

- In general, this is any solid - made of plasticor glass, metal or plexiglass and other materials. That is, something dense. If the subject is not too soft and can vibrate, it is suitable for work. The shape and size of the object can also be any: flat or complex, very small or large. The number of sensors depends on the size of the subject and the intensity of the interactions that we track. The more events, the more sensors are needed. For example, we trained a car door and used only three sensors for this, and you can control the process within up to 40 actions. Tap in different places, brush away, turn the handle, open a window or door.

“But vibration analysis requires a lot of resources, how did you manage to get by the integrated chip without using cloud computing?”

- The advantage of HyperSurfaces is that everythingoccurs in real time, and the delay does not exceed a few milliseconds. This requires working on embedded circuits without an internet connection or cloud technology. So that it can be used in the middle of the desert, if necessary.

Photo: personal archive

Therefore, we are exploring opportunities to reducesizes of neural networks and other algorithms based on built-in sensors. Many have to sacrifice. But we need everything to work as fast as possible, so we are working on optimization.

- Your first project Mogees translated vibration from any object into music. Are these technologies related?

- Sure. Mogees was the beginning, there I found inspiration. He began to plunge into the world of signal processing using AI. Rethought, re-imagined the objects around us. Projects are certainly connected by inspiration. Some technical aspects of signal processing and vibration. After the success of the Mogees, we worked to ensure that technology became available not only to musicians. We were looking for the opportunity to translate vibration not only into melody, but also into knowledge or command. Give access to any business and manufacturer. Thus was born HyperSurfaces.

Why do universities need intellectual property for student ideas

- Many young scientists complain that it is very difficult to transfer ideas from the scientific community to real business. How was it in your case?

- Very hard. With Mogees, we were lucky that the product itself was very clear. We posted some simple videos online and they caught the attention of a large audience. Therefore, the university also got an idea. Not that professors do not want to help, they are very supportive. But the system is out of date.

Now a lot has changed, but when I was gettingfive years ago, all intellectual property rights belonged to the university. It's hard to make a startup with such a slow investor. Now schools have become smarter, giving students the choice and support.

Photo: personal archive

In startups, it's important to be mobile and acceptsolutions instantly. This is not possible if decisions are left to universities. Therefore, it is so difficult to raise money: you tell your partners that you have to wait a week or two, and they don’t listen. The market is full of other offers.

Now the situation is changing. In any case, in the United Kingdom. I know that Oxford, Cambridge and Imperial College have simplified the conditions for students who enter into venture projects.

Oxford, Cambridge and Imperial College have created their own venture capital funds, Imperial Innovations, Oxford Sciences Innovation and Cambridge Innovation Capital, providing students with opportunities:

  1. Patent your research.
  2. Create a company based on developed innovations.
  3. Licens technology for partners in exchange for royalties.
  4. Assist in attracting research grants.
  5. Sell ​​research materials through Quicktech, a direct online licensing platform.

“But sometimes universities try to take copyright over student inventions?”

- This is a confusing question. In my personal experience, many professors or officials do not know at all what intellectual property is. Or the student publishes the discovery, and, obviously, it is not intellectual property, because it is in the open source. Or leaves this knowledge to himself to open a startup.

Students rarely get copyrightuntil the professor asks for a patent. But in this case, the rights may remain with the university. Everything that is published is not protected by copyright: a dissertation, an article in a scientific journal are open sources.

- The Mogees project successfully raised funds for Kickstarter. How important is the simplicity of the invention to successful crowdfunding? And what about HyperSurfaces in this regard?

- Yes, most of the success of the Mogees was thatIt was made easy and understandable. We focused on children interested in music, authors, artists. HyperSurfaces works with designers, developers, and manufacturers. It's okay that the technology is a little more complicated, and you can figure it out not in a few minutes, but in a couple of hours. More precisely, on the current day, it's even a few days, we are improving the product. Working on it. We are creating a set of developers so that it becomes as simple as possible for customers to work with technology.

Why controlling the Internet of things is getting harder

- Can HyperSurfaces become a tool for total control, taking into account the convenience of integration with artificial intelligence systems?

- Any technology can be used for the good or the detriment. It is important for everyone who invents to talk about it, to control that technology works only for the benefit of humanity.

I emphasize, we work with technology"Artificial intelligence on EDGE", which involves working with data locally, directly on the device, and does not require an Internet connection. This helps to avoid the threats of data leakage that are unavoidable when using cloud computing, where you cannot fully control who has access to the data. The design of cloud technology itself implies that for work you have to transfer data to the network, accordingly, it is very difficult to track where they went.

- But this is regulated by privacy policies - for example, GDPR.

- Yes it is. That is, regulation is necessary. Just what the European Union developed. More transparent personal data access policy. The EU is doing a great job, improving the security situation of personal data, and I hope that other countries will also follow this example.

Photo: personal archive

But in our case, the technology itself is made in such a waythat work with data occurs locally. The developer decides whether to share data with external sources. For example, you make cars. You use our technology without a network connection and you can guarantee customers that the data will not go anywhere.

“But will it not become more difficult in the future to control this possibility of choice?”

- Oh sure. It will be trillions of IoT devices. They will be in our clothes, walls and floor. Everywhere. It does not matter, with or without AI, the question of who will manage this data and have access to it remains important and should be widely discussed now right now.

- Previously, touch screens understood keyboard commands, then they learned to understand gestures. What new gestures will be required to use HyperSurfaces technology?

- Good question! I can give a concrete example of a gesture, but in fact the question is much wider. I see this technology as a complete paradigm shift. We give the designer a new tool to design interaction in a new way and to be able to recognize any event that has vibration.

It’s akin to before the invention of mobileNo one needed a phone or applications. But its creation made it possible for a new generation of designers to completely rebuild the world around us, to change the usual interaction with it. So it’s even hard for me to imagine what ideas will come from a new generation of thinkers and developers. They, in turn, will reinvent the world with HyperSurfaces.

This will affect many areas: cars, home appliances, electronic devices, IoT and security. Technology for those who are currently developing products for their good ideas. You can imagine the whole "Hyperles" that gets in touch. He can warn of the danger that threatens. For example, what is happening now with the forests of the Amazon. This technology offers endless possibilities. And the task of users is to reveal them.