Quantum Lab Poland receives Horizon 2020 prize
Few areas of computer science stir up visions of a future straight out of science fiction than machines that know how you feel. And yet, the future is now, according to Polish start-ups such as Quantum Lab Poland and their emotion-reading software. The European Commission agrees with them, as they are the most recent recipients of the prestigious Horizon 2020 prize
The Horizon 2020 prize is part of “the biggest EU Research and Innovation programme ever” according to the European Commission website, and its goals include giving financial support to start-ups and technology companies conducting cutting edge research in Europe. Quantum Lab was one of three Polish companies to receive the prize this year (the others being NapiFeryn BioTech and Atsora).
I spoke to Konrad Bocian, the Chief Science Officer (CSO) at Quantum Lab and a Ph.D. candidate at SWPS University, to find out more about the start-up and what they plan to do with their new prize money.
Web.gov.pl: Tell me a little about the history of Quantum Lab – when was it founded and by whom?
Konrad Bocian: Quantum Lab was founded in 2013 by four guys: Bartosz Rychlicki, Krzysztof Cywiński, Paweł Tylingo and Damian Jagielski. Our company consists of Quantum Lab Poland, which is an R&D department in Gdansk, and Quantum Lab Co, which is our headquarters in the U.S, in San Francisco. We also have a sales department in Sweden. So far Quantum Lab Co. has raised $350,000 from the first seed round and $130,000 from tech grants. The principal strategic company investors are the investment funds Black Pearls VC, Business Angel Seedfund and Profound Ventures.
And what is your role in the company?
I joined as the fifth member of the team. I came on board as a psychologist and am a 2nd year Ph.D. student in psychology. When I met Bartosz and the other team members, we quickly discovered that we all share the same idea of combining psychology with advanced technology, and that was the starting point for our first project.
What are some of the projects that QL is working on right now?
Our mission is to teach computers how people feel and right now we know the best way to achieve that. Bartosz, our CEO, is a “psychological freak”. I say that because he’s an entrepreneur and an IT guy, but he also loves to learn as much as he can about results from psychological research. He came up with the idea that it would be cool to help people overcome their own limitations with the combination of psychology and mobile technologies.
The first idea was to make an app that would help people to be more punctual, because Bartosz has a problem with waking up on time. That was the first project, but it wasn’t accomplished, so then we moved on to mood disorders. We thought that we could make an app that would help people monitor their moods. The idea behind it was that if you can track something, you can make it better. The same idea is behind most fitness or workout apps - monitoring your progress.
That’s how the How Are You App was born. We built a mood tracker that right now we are selling worldwide. It helps people monitor their moods. It’s designed for people with mood disorders or depression, which is a huge problem right now in developed countries.
When it comes to psychological help, often it’s very expensive, and a lot of people can’t afford it. That’s why we called the How Are You App “Therapy in Your Pocket”. It’s made to be the first line of defense against depression.
We spent about half a year developing the app. It came out in English on iOS first and then on Android. In the next two months we translated the app into Polish and also Spanish. After almost two years, the app began gaining traction, and right now the sales are better than before – but you cannot establish your company on just one app, even though depression is a huge problem worldwide. Accordingly, we and our investors decided that we have to move forward and think about something else, another product.
So what else has Quantum Lab produced?
Some of our potential clients suggested that it would be interesting to evaluate peoples’ emotions and moods in a non-subjective way, because right now the How Are You app is more like a test, where you answer questions and then the algorithm combines the results. The new idea was to figure out how to track emotions and recognize the moods of people without asking them.
This is how we got into affective computing, which is a part of computer science that combines IT with psychology and is about teaching computers or social robots how people feel and how they show emotions. Affective computing helps scientists and engineers to teach social robots how to recognize and emulate peoples’ emotions. So we decided to build an engine to do the same thing.
And we did it in about a year and half. We developed an algorithm that successfully recognizes emotions. Once we had a ready to go engine, which we called the Xpress Engine, we thought about the product – how could we sell our engine, where could we use it? We thought about market research, because looking at our competitors, we noticed there are companies out there who are successfully selling products that help recognize emotions in the market research area. Specifically, two companies: Emotient and Affectiva, both U.S. companies that are spin-offs from MIT. Both also had investors, so we knew there was potential to build something even better and more effective. Since there were no competitors in Poland, we thought we’d build the product for the Polish market, and then move on to Europe.
The video shows how the Xpress Engine software “reads” different facial expressions and registers them as emotions.
How exactly do you teach a computer emotions?
Dr. Paul Ekman, one of the advisors to Emotient, published an article in the 1970s about facial coding (FACS). He described the muscles of our face to teach people how to recognize how they move and what they look like when they’re active. Because of that, we can teach computers the same knowledge. We can tell a machine that if I’m using Action Unit 12 (the zygomaticus major muscle), for example, it means that I’m smiling, and the emotion reflected by that is happiness. Dr. Ekman described six basic emotions as universal, and independent of culture - the only differences are how much of them we show and in what context.
Based on this knowledge and knowledge from affective computing, we taught our algorithm how to recognize emotions, then we built a product that we call Ellen.
What does Ellen do?
Ellen is a platform for market research. It allows market researchers to conduct fast and efficient research on consumer emotions. For example, if researcher has materials such as an advertisement, they can upload this material to our platform, then we generate a link and they are able to send it to a panel of respondents. The participants can then click the link and the program launches. The participants go through a tutorial about how to sit and lighting and so on, then they watch the material, and we record their faces and send this recording to our servers, where the Xpress Engine goes through it and establishes the possibility that there’s an emotion on a face and determines what emotion it is.
Based on that, we can compute the data for the entire sample; for example, we can study 300 participants and tell our client that in the 10th second of the film material, 70% of the participants were smiling or were sad or angry. Based on that, we can draw particular and reliable conclusions for market researchers or advertisers.
The video shows a person watching a video advertisement and their reactions being recorded by the software. The software is then shown to evaluate their emotions based on their facial expressions.
Have you had any commercial success with Ellen?
Right now we are trying to sell our product on the Polish market, but it’s not easy. I have the feeling that most of the Polish research market is not keen on innovations. They tell us it’s very interesting, but they have a hard time understanding the value of our technology. We understand these issues and appreciate feedback from our clients, which helps us better develop the product. The first version of Ellen was simple and basic, but right now it’s a comprehensive tool for market research thanks to our clients, who requested new improvements. We develop Ellen each day, adding more and more features that help our clients conduct research rich in emotional data about the feelings that are triggered by their ads. Right now we have over 20 happy clients (including Play, Lotos, PBS, TVP and TNS).
Our main aim is to make Ellen fully automatic. We want to make a platform to which our client can log in, choose how many participants he or she wants, from which part of the country, which gender, age, etc., load the material, pay by card, send the links, and obtain results in less than 24 hours. There’s no such thing on the web right now.
What is Quantum Lab doing with the Xpress Engine?
Right now we’re thinking of developing our Xpress Engine for other areas, such as medicine. We want to help people who are paralyzed. After talking with several doctors, we learned that when it comes to people who are paralyzed, a huge issue for them is communication. Right now, people who are paralyzed can use eye tracking, which means they can use their eyes to move a mouse cursor. The problem is that in Poland, this product costs at least 45,000 PLN, which is a huge challenge to people receiving 800 PLN per month of social support.
We realized that we have a simple product that might help, because when it comes to Ellen, all you need is a webcam and an Internet connection. If we use the Xpress Engine, the software behind Ellen, we can develop a ready to go product. For example, a person who is paralyzed can use a webcam to communicate with their family members or doctors in an emergency situation. We want to teach our program to recognize the meanings of certain actions by paralyzed people, such as blinking their eye, etc. We call it the Xpress Helper.
I think the huge advantage of our product is that it could cost about 1,000 PLN, so basically a webcam and our program, and we can connect our program to a smartphone as well. We could also develop a mobile app for the family members of a paralyzed person so they can monitor what’s happening with their family member, and they can communicate with their family.
This is one of our ideas. Another is to help develop the software for the social robots that are being built in technology universities in Poland, to help them recognize the emotions of people.
How did you come to receive the Horizon 2020 prize? Did you apply or did the European Commission find you?
We received the prize because of Ellen and the Ateknea Solutions company, which helped us wrote the grant proposal. We applied [but did not win] three times with Quantum Lab Poland, and it was the fourth time that we received the prize. That’s when we had a ready and comprehensive product with some traction from the market. The reviewers recognized that we had a product that was ready to go abroad, that was well developed and reliable. When it comes to the evaluation, the threshold is set very high. In the end, we needed a year and half to make a product that could win this prize.
What are Quantum Lab’s plans for the prize money?
Most of the money will go towards Ellen, and to trying to conquer the market research market in Europe. It’s a huge boost for our company, because we can further develop our product and actually try to compete with start-ups from the U.S. that are well-funded, while we are a small start-up from Gdańsk.
This is the first phase, which is 50,000 euros. Thanks to that money, we can conduct a feasibility study, put some cash into marketing and make the product more visible for potential clients. Furthermore, we can conduct case studies and benchmarks, which are important for our clients
After the first phase, which is crucial, the reviewers will decide if we should receive the second phase, which is 1.5 million euros. This would give a huge boost to our company, and our products as well. More importantly, it will help motivate people to work even harder to bring our clients the finest possible product. With money from Horizon 2020 we would be able to develop not only Ellen but also the Xpress Helper, and maybe - this is our dream - to build our own social robot. Right now, it’s only a dream - we’ll see if that’s possible in the future.
Editor / Journalist
Got any questions? Send an e-mail to the author