Skandi style via algorithm AI helps set up

Skandi style via algorithm AI helps set up

Whether it’s a round table in the dining room, a colorful picture over the sofa, or a larger lamp in the hallway – augmented reality apps can help with interior design. However, much more will be possible using artificial intelligence in the future, explains Ilija Vukorep, architect and professor at the Brandenburg Technical University Cottbus-Senftenberg (BTU) in the field of computer-aided design in an interview with n-tv.de.

Has the trend topic of artificial intelligence already arrived in the home furnishings?

Ilija Vukorep : Yes, very much. For example, what do you do first, if you want to rebuild? They look at pictures in the social network Pinterest. This area of ​​artificial intelligence, image recognition, has had the strongest impact on interior design. But the topic is big and far from grazed.

What does that mean?

Pinterest is not specialized in setting up, but simply a picture cataloging machine. Soon, however, platforms will be launched that focus on the subject of living. These companies collect all sorts of interior design images, feed the computer and try to draw conclusions from it. Many companies are working on this, but there are no finished products yet.

How will they work then?

Above all, they will work with the transferability of styles – for example Scandinavian, rustic or hypermodern. Users can upload photos or floor plans of their home or rooms, and then generate styles in the form of images. This technique is called GAN (Generative Adversarial Network). The second stage will be to turn a style into concrete plans – where exactly a chair could stand and where to buy it. A budget could also be set. Stage three could be the continuation of the 3D printing of a piece of furniture or furniture.

There are already apps with which I can move furniture and experiment with the device.

These are augmented reality apps that let me see in 3D what it looks like when my shelf or picture hangs at a certain height. However, this is still awkward, and often only a single product can be visualized. I hardly know anyone who uses this. These applications need to be matured.

How detailed will an artificial intelligence set up a room?

Curtains, pictures, bedside lamp, candlesticks – that’s all part of it. That will also be the claim of the users.

Will the technology be used more in the commercial than in the private sector – more by interior designers than by private individuals?

Both will develop in parallel, and both groups will use similar tools – professionals only in a more complex form. Architects and designers can help artificial intelligence when many rooms need to be designed and quickly furnished. A human would plan each room one after another. The machine that does that with machine learning unwinds all the rooms at the same time. For example, the US company WeWork, which converts and leases office space to coworking spaces worldwide, uses artificial intelligence to respond to different cultures and ways of working, so that a coworking space in Singapore looks different from Berlin.

When do you expect everything you describe?

Everything is still in the making. Maybe it will be normal for us to use such apps in five years, but not in two years. The price is also a problem: As soon as an app costs something, it is difficult, because the user actually wants to pay nothing for it. Since you have to see how it can be used to make money.

Do you see any other problems?

Yes, the danger of the so-called bubble. Currently, artificial intelligence specializes in following the wishes of users. That is, it reinforces them in their opinion because they only see things that suit their tastes. Only when we can add something like a critical voice with alternatives to the software, we really have an intelligent system.

They also do research on computer-aided design in building. How can artificial intelligence be used there?

Among other things, my team and I are involved with autonomous construction and research on rope-moving robots. The goal is to use learning algorithms to hover the above-ground robot on construction sites or in prefabricated halls. There, he would adapt to constantly changing working conditions on a construction site, such as the growing construction, equipment and people in the workroom or the weather, independently. But it will be some time before robots build houses themselves. Building is a very complex process, which is usually done by a team of experts from different disciplines. An artificial intelligence can not simply take over completely.