Mr. Zent, what is the secret of a good computer game?
A good game has a healthy mix of challenges and rewards and gives gamers a sense of wellbeing. But until now there was no way of measuring those effects. Developers have limited ability to predict the user experience. To somehow address the problem, they invite testers to play the game, but their feedback is limited to the pen-and-paper surveys or talks with the developers after the test. All of this is extremely subjective.
What does this mean for the development of new games?
Despite gigantic development costs, many games flop in the market due to a poor user experience. To improve the user experience, developers need objective data that show exactly what players feel and how engaged they are at any moment and location of the game. Developers could then use this information to optimize their products.
These are the data that your startup Brainsight will deliver. What technologies does it take to objectively capture testers’ emotions?
Brainsight uses an Electroencephalogram, in short EEG, to measure the brain activity of the testers while they are playing the game. Then we apply machine learning to analyse the data. First the algorithm is trained with the labelled EEG data of people watching emotional videos. From this data the AI learns to predict the emotions from the brain activity. This way we can determine how focused the person is and if he or she is experiencing positive or negative emotions.
So this technology can show what is happening in the gamers’ heads?
We don’t know what they are thinking, but we can measure which regions of the brain get activated and how strongly. From this information the trained AI can figure out which types of emotions are present in the game. For game developers, this technology opens up entirely new possibilities. For example, they can now see the exact moment where a tester starts losing focus and gets bored.
How did you get the idea for AI-based game testing?
It came from TUM, actually. Vlad took part in the BCI & Neurotechnology Spring School 2021 because he is fascinated with the brain-computer interface – BCI for short. One of the invited speakers was a researcher who talked about many potential applications of BCI in the gaming industry.
So you decided to create a startup?
Vlad and I had met a few weeks before during the virtual application process for the Think.Make.Start workshop. We were looking for a business case for a neurotechnology startup. As passionate gamers, we quickly agreed that using neurotechnology methods in the playtesting phase could help developers to improve their games. That’s because there’s one thing we both agree on: we have no time for bad games.
How much progress have you made with this new AI-based technology for detecting emotions?
The algorithms have learned how to interpret EEG data. We can use AI to determine gamers’ activity and emotions while they are playing. These values can then be mapped to every moment and location of the game. We will now offer this technology to game developers.
Are there other applications for this emotion detection method apart from playtesting?
Neurotechnology has enormous potential. It’s basically about networking the human brain with computers. It’s still early days. Recognizing emotions with EEG data is just a small step, although it could benefit a lot of users.
Can you give us some examples?
Film producers can find out whether a certain shot can arouse the intended feelings in the audience. Music streaming services can test whether their playlists create certain moods. The technology is also interesting for marketing agencies. They can see if their campaigns trigger the desired emotions.
Think.Make.Start offers the opportunity to develop innovative business ideas and prototypes in just 10 days. How much did you benefit from it?
Of course, it was great that we met there. But then the pandemic hit and the program was postponed until the fall. So, in the meantime we kept working by ourselves to develop our business idea. In the end we won the Start Award for the business idea with the most potential.
At the moment I’m taking part in a business plan seminar at TUM where I’ll do the analysis to decide which industry will be our next expansion target. With the support of Prof. Cheng of the Institute for Cognitive Systems we also applied for the EXIST Business Startup Grant this month. If that works out, we can work full-time on Brainsight starting in February 2022. We’ve already found first customers.
And you can do all that and still keep up with your studies?
Surprisingly it’s no problem. TUM actively supports students who want to start companies: You actually get credit points for a lot of the workshops and courses on entrepreneurship such as Think.Make.Start. And if everything goes according to plan, my master’s thesis will be the Brainsight business plan. So, you can complete your degree while working on your startup – which is very cool.
- Philipp Zent is studying Management and Technology at TUM. Vladislav Samoilov completed his master’s degree in theoretical particle physics at TUM in 2021.
- For the development of their business ideas they received the Start Award for the business idea with the greatest potential in the 2021 Think.Make.Start program.
- Brainsight became third in the TUM IDEAward 2021 at the end of November.