“If you don’t have 10,000 good data points, you get nowhere”: Intuiface’s Vincent Encontre on Machine Learning.

One of the hot topics in recent years is machine learning and AI (Artificial Intelligence). You can see too much of it in the promos from companies beyond the IT industry. You might righteously ask, as a customer, what does it mean and where’s the catch? During our podcast interview with Vincent Encontre (COO of Intuiface) at the DSSE earlier this month, we discussed the prospects and challenges for interactive screens and machine learning technologies. To address this topic, we asked Vincent: Do you use machine learning in your interfaces and your technology and on what level?

“First of all, you’ve mentioned two terms: AI and machine learning. And in fact, you’re perfectly right – we should be talking about machine learning. The rest is just a buzzword. We are using machine learning in only one place, which is about the business we’re in. When a user wants to explore an API, then we select the response from the API which can be complex even for the developer. We want to simplify that view and offer a structured view to a designer at a designer level. So we’re able to set a query: well that's an image, so let’s show that’s an image, or that’s a URL, or text or description. So we’re using machine learning here to discover what might be interesting in the API on a design standpoint.

Again our customers are collecting data information that goes into analytics, but there’s one per usage. So for example, we’re in no way to discover (maybe we can, but that’s a long stretch) which icecream perfume will have more success because that day the temperature was higher than usual and these kinds of things. We can collect that. Again, I assume we can import the current weather conditions in experience so our customer back in Italy has put together when a user is selecting this icecream with that perfume the weather conditions were these: cloudy, but 34 degrees Celsius. We can advise the user to correct the right information, but it will be up to them after enough data points corrected. Even if machine learning has made a lot of progress, if you don’t have 10,000 data point, good data points, you get nowhere. So maybe after that, they will be able to use machine learning, but on their data set, not on our data set.”

http://gph.is/2oQT5M2

Listen to the full podcast interview with Vincent Encontre to learn how interactive digital signage creates new marketing possibilities and how machine learning can pave a way for a regular business to succeed:

Kitcast Podcast feat Intuiface – Episode 6 – What About Interactive Digital Signage?