‘’Why choose between expert knowledge, machine learning, and deep learning. As long as you have a proper fusion platform available, you can have all three team-up for you each working their strengths in concert’’. And our next speaker from M_AI summit 2018 has already explained how to do that.
Meet Ralph Humberg with the topic of Tama’s AI Avengers.
Ralph founded the Tama Group in 2015, as a spin-off from his own technology platform 2D-Ventures. Ralph’s background is pattern recognition from TU Munich and the Tokyo Institute of Technology. In his more than 20 years in the industry, he held various leadership and senior management positions in Trimble, Definiens, ADVA Optical Networking.
He explained that they deal mostly with images, in general, anything that the human eye can see. They use, combine, fuse them and do a lot what they think is possible with a machine.
There are three topics he touched upon during his talk.
- AI team play vs pure play in imagery
- The question prior to coding
- Business examples in construction and remote sensing
‘’We certainly do deep learning but not exclusively we actually use a lot of classic methods and in combination because we find it a lot more effective, straightforward and leading to proper answers. And actually, you won’t see any consumer applications. Our business is completely B2B. And what we do or always what our interest is that we want to help decision makers. So that’s my first key statement. We are not here or Tama group is not here to replace humans at all. What we build is what we call decision support systems. We are interested in defense work we don’t do really anything on the consumer surveillance stuff. All we do is to help experts become smarter with their decisions on what we call assets’’ said Humberg.
The Tama group is focused on AI primarily for B-B and B-G applications for persistent assets. Their technology allows asset experts to take better decisions by deploying their decision support platform.
‘’We talk about images so we all see images and the beautiful and the difficult thing is each of our brains actually can choose the world in his own way, right? My perception of the world and the one of any other guy is completely partly different and to billions of people on the planet that’s different. So we’re working on it for years and how do we get actual access to tap into these brains. There are different methods doing that and well I’m in that now for almost 20 years, and we’ve tried several methods and we still keep on actually doing them in parallel because we think some work better than others depending on what it is. So the first way we did is we conducted a proper assessment. So a smart person talked to maybe 10 experts and tried to understand what do you see in that particular image with a stack of images. And after that, we wrote the problem assessment down and had a code start doing system. So we had a system design so far. The only problem is that the one who’s doing the coding never really had access to the brain of this guy, the person who really understands what he’s doing’’ explained our speaker.
He further dives deep into the topic explaining in detail how actually they build their work.
This was just a snippet from Ralph Humberg’s talk. Sounds interesting? You can visit Develandoo AI Innovation Lab‘s youtube channel and check out the full video of his talk if you also want to dive deep into the topic. In our channel, you can also find videos of all our speakers’ talks.
We are thankful to all our sponsors and participants and the speakers for attending our Munich-AI Summit 2018, the one and only free event at Munich related to Artificial Intelligence. We invite to our AI platform all representatives in the field of Artificial Intelligence, AI experts, and simply those who are interested in the exchange of ideas within AI innovation, and big data analytics in 2019.
- Artificial Intelligence