Funartech, the startup that strives for flexibility within artificial intelligence

Funartech, the startup that strives for flexibility within artificial intelligence

by Startup Montréal
27 September 2019

Do Artificial Intelligence as we do – in the form of hybridization!

– Nikolaj Van Omme, CEO and co-founder

Funartech is a young Montreal company firmly rooted in Artificial Intelligence, but not only that: this startup is working on “the AI ​​of tomorrow”. Founded by Nikolaj Van Omme and Dania El-Khechen, Funartech’s goal is to solve industrial problems through ML (Machine Learning) and OR (Operations Research) hybridization. A path different from others, complex, but worth it, according to Nikolaj Van Omme.

Funartech spoke to Bonjour Startup Montreal’s team about its mission and AI’s future, during the AI: Big Buzz Event at the Pop-up Lab.


How long has Funartech been around?

We’ve been around for almost 2 years, we were incorporated on October 23, 2017.

Tell us about your mission.

We want to use current technologies to improve our environment and have a positive impact socially and beyond. Why not use our technology to do good in our society, especially for our environment? That’s what motivates us – we’re consultants first and foremost.

We have found a new way to hybridize the OR (Operations Research) and the ML (Machine Learning). We have interesting results that are worthy of exposure. We are a service company, but we are moving towards a transition to also become a product company.

In concrete terms, what does it mean?

There are basically 4 ways to combine machine learning and operational research:

First of all as two independent black boxes: we first us ML and then we optimize on these predictions with OR. Here is my favorite example: you are a train company and you have to replace or repair your tracks. How can we do it at a minimum cost? Use the ML and then the OR! First, you put cameras under your wagons and take pictures of the rails while your trains roll. With Deep Learning, you reassemble photos and detect defects along the rails. Once you’ve done that, you have a map of your entire network: you know when and where to intervene but you do not know how. This is where the OR comes to play. Indeed, it is possible that if you assign a given rail to a repair team, it would be cheaper to use the same team to replace it, but it wouldn’t be done within a year. Therefore, the OR approach would be cheaper than sending a team to do this replacement a year later. This is finally beginning to be recognized and used. Especially since… 2019!

The following two ways are technical. Let’s just say that we can use one to improve the other and it works very well. The first results are very promising.

Finally, and that’s what we do at Funartech, a complete hybridization. We developed new algorithms. To give an idea of ​​the power of this new hybridization, I will give two examples. The first is the injection of emotions into IA. I’m not talking about detecting emotions in users but to make the robots emotional. Our system not only allows to change emotions but also to be content with few data and it makes decisions autonomously. The second example goes even further: this is like the first algorithm but “more intelligent”. Indeed, we discovered how to develop an autonomous system that is able to relearn its environment at a point while having erased all that has been given to him beforehand. It’s a form of intelligence and we call it “an open door to the AGI”. AGI means Artificial General Intelligence and it’s the “real” intelligence, which is measured and will eventually exceed human intelligence. Through this, we support our customers by improving their processes, whatever they are, to avoid expenses – or maximize their income.

Have you received support from the ecosystem?

The AI ​​Launch Lab has been helping us since the beginning of the year. We proposed a project for their January cohort and then for their May cohort. Next year, we will do research with Professor Michel Gendreau, who believes in us and in the hybridization of  OR and ML. It’s an incredible opportunity.

What are the key resources of Montreal according to you?

Montreal has an extraordinary asset: it is the presence of a highly respected operations research community worldwide – we are leaders in the field, and this has been for 30 years, if not more! Montreal also has great expertise in AI that is internationally recognized. There is only one other city that has this asset: New York!

Do you give back to the startup ecosystem or other founders? If so, in what form? If not, is it in your plans?

Ultimately, we would like to be able to give back to the ecosystem by investing in smaller companies that do not have access to funding or resources – just like us right now!

How could the Montreal startup community learn about other ecosystems?

We are part of these companies that are in a gray area, we meet some criteria but not all, so we do not have access to resources that could help us. It would be interesting to look into it, not only for us but also for all others! We are in a gray area because we are too much ahead of our time. Our expertise is not recognized – not because it is not effective but because it is unknown, even though we have more and more traction. So, it’s the game of the chicken and the egg. In order to demonstrate the effectiveness of our approach, we need a big project but to have a big project we need to demonstrate the effectiveness of our approach. Then, how can we get recognized for doing AI differently, and I mean really differently? For companies like ours, and we are not the only ones dealing with this dilemma, the choice is simple: you have to climb the ladder, little by little, proving at every step the merits of our approach.

You held an event on September 11, can you tell us more?

The goal was to explain who we are, and why we do what we do.

In addition to that, we are launching our Bot project, focused on current environmental issues. It is provided with our new technology that mixes emotion with artificial intelligence, in order to help with eco-anxiety. Eco-anxiety is a form of existential distress caused by environmental change, an increasingly widespread phenomenon. This Bot is interactive and pushes for action. Finally, it is also able to provide information globally but also locally.

The event was the result of a partnership with Collab Machine, an independent collective whose main idea is to bring together the strengths of different talents to lead to the success of selected projects.

Final word: Do you have any advice for entrepreneurs who want to get started in AI?

Do AI as we do – in the form of hybridization! We think that’s where the future of AI lies. The results obtained so far are promising and go in that direction. By choosing hybridization, you will go further, and in the right direction. Beyond that, we need to help each other!

There are a plethora of innovative projects in Montreal all the more incredible ones than others. It goes from new energies to generate happiness at work through the cleansing of air and water-based on seaweed. Projects are very risky but the results always lead to a very positive feedback for the whole community. If we succeed, we are committed to investing in this type of projects and we are doing so for the good of the community.