LØRN case C0051 -

Davide Roverso

Chief Analytics Officer


«Virtuelle maskiner»

I denne episoden av #LØRN snakker Torgeir med Chief Analytics Officer i eSmart Systems, Davide Roverso, om Big Data og hvilke intelligente tjenester det åpner opp for. eSmart Systems utvikler primært sky-basert og AI-basert programvare for nett- og energiselskaper, og i samtalen med Torgeir deler Davide sine tanker om både de spennende mulighetene og de etiske utfordringene Big Data utgjør, samt interessante fakta og tall om teknikken som ligger bak denne teknologien.
LØRN case C0051 -

Davide Roverso

Chief Analytics Officer


«Virtuelle maskiner»

I denne episoden av #LØRN snakker Torgeir med Chief Analytics Officer i eSmart Systems, Davide Roverso, om Big Data og hvilke intelligente tjenester det åpner opp for. eSmart Systems utvikler primært sky-basert og AI-basert programvare for nett- og energiselskaper, og i samtalen med Torgeir deler Davide sine tanker om både de spennende mulighetene og de etiske utfordringene Big Data utgjør, samt interessante fakta og tall om teknikken som ligger bak denne teknologien.

17 min

Choose your preferred format

Velkommen til Lørn.Tech – en læringsdugnad om teknologi og samfunn med Silvija Seres og venner.

TM: Hei, og velkommen til Lørn.Tech. Jeg heter Torgeir Mikaelsen. And together with me… we will do this episode in english because… David Roverso, welcome!

DR: Thank you.

TM You have some, obviously you're not from Norway originally

DR: No. From Italy.

TM: From Italy. Fantastic.

DR: I have been in Norway for a very long time. So.

TM: Exactly. You work in a company called E Smart Systems.

DR: Yes.

TM: And you do have a lot of clients in Norway, but you also work internationally. So could you start by telling a little bit about the company and where you work today?

DR: Yeah, it is a software company. It's not a startup anymore now. Now it´s a scaleup. We are about 75 People based in Halden, but we have offices in Oslo in London, and we're expanding a lot in the US and also in Denmark and Germany. We developed software, mostly cloud-based software for energy companies mostly electrical utilities. So we work in the power and utilities sector. So the core the company is a bunch of people that have worked in this domain for a very long time for 30 years and kind, of about five years ago, they start hearing about all this new technology about Big Data and Ai and they took a trip to the US to Silicon Valley where they were for a few months and then came back home with the idea of, yeah, how can we apply this new technology to this domain of power and utilities. So that's how the company started and since the beginning it has had a very big focus on both Big Data and Artificial Intelligence. And that's kind of my role in the company, I'm Chief analytics officer so I'm responsible for all that we do around analytics and AI.

TM: So. But I can say it so you don't have to say to yourself. You're an academic.

DR: Yeah. Yeah, I originally.

TM: Yeah exactly. So you actually know a lot about this phenomen called Big Data and Artificial Intelligence. Not just a guy running around in conferences pretending to know!

DR: Yeah. Now everybody's is talking about machine learning, AI and Big Data.

TM: Exactly.

DR: Not too many have had a long experience in that area and a lot of people kind of rebranding themselves as data scientists when they have different type of backgrounds because of course there's a huge demand for this for this competence. Also a lot of people are actually learning new skills now a days. Is is mutch easier than, I mean, you don't need to take a PhD in machine learning to be able to do data scienc. People follow courses on the internet and that's very helpful. And there are a lot of very bright young guys, and some of them are with us. So we are very proud of our team.

TM: You should be, you should be. You said that in E Smart System… you have obviously lots of experience in data analytics, Big Data, Artificial intelligence, but you have like a domain knowledge about the energy sector in Norway. Is that still one of the largest client base for the company, or?

DR: Yeah. It is. We kind of think about this three domains that we need to have competence in. One is more the data engineering part about Big Data and Cloud systems and that kind of thing. The other part is more on the AI and machine learning. And the third one is the domain knowledge on the area in which we apply this technology. So power and utilities. So we have a lot of people with that kind of background on this one. So we think that the combination of these three is kind of the core of what we do, and it's kind of our advantage in this area.

TM: So let's talk about Big Data and continue to talk about the energy sector. I think everyone agrees that transports in electricity, power utility station, that those installations creates data. But what is actually the data? Could you give people some examples when you collect collect data from the grid? What do we actually collect?

DR: Well, before we go into that, I think one interesting fact is that many people don't think about it in that way. But the power grid is actually the biggest machine ever made and the most complex machine ever made. You know, you have hundreds of thousands of kilometers in hundreds of thousands of components and everything needs to be synchronized together in order for - even me - to be able to charge phones and watch TV.

TM: Exactly. In my old life in politics we used to…web developed an expression and how to chew… to visualzie how big it is! So that, we created, like - it's the greatest the highway for electrical energy and the most complex highway because, yeah as you say. How many kilometers did you say? Thousands of kilometers only in Norway?

DR: Yeah. 300,000 kilometers just in Norway. You know, the greatest is the changing and this last year's there has been a lot, you know. We're going more towards the smart grid so originally, you know, power distribution was kind of a centralized thing. You have a big power plant and then you transmit energy over long distances and then you distribute it to the lower voltages to the houses and so forward. But it was just a kind of a one-way thing and not much data was collected in that time. So a lot of this utilities, electrical utility, we are kind of operating blind because they were not measuring much of what was going on. Especially the lower voltage grid.

TM: So, a small lover voltage grid is kind of…it's dumb?

DR: Yes. Yeah, but now it’s a lot of people in Norway who have experience. Most of the people of getting smart-meters at home. So then suddenly electrical companies have instruments in every house, you know, so they can directly measure with higher frequency consumption and all other parameters. So there is much more data. Data that they need to be able to handle, and also there are many more opportunities to be able to operate the gradius in a more optimal way. Because, the grid is changing not just because we are instrumenting it, but also because we have what we call distributed energy resources. Or people who are installing solar panels. There are windmills. There are electric cars that needs charging. So it is a much more complex and flexible grid that you can’t operate blind so you need data.

TM: Exactly.

DR: So that's one big source of data. Nowadays there are many more measuring points and much more data that comes in that needs to be analyzed. So that's one area we work in, kind of using machine learning to make predictions of what kind of load we can expect in a certain area or certain transformer station and so forward. Another big area that were working on is more related to inspections. We are…we said that the power grid is kind of of the biggest machine ever built, but was guilt quite a long time ago.

TM: So a lot of them is getting old?

DR: Yes.

TM: So in need of new adjustements.

DR: A lot of infrastructure is is is quite old. Some is hundred-year-old. And if you look at the number of outages, I have some figures from the US just in date, is that they had kind of five major outages every year. Nowadays there are above 200 every year. So there is an increased need for the for inspections and maintenance. And in that area we are using AI to analyze the big amounts of data that electrical companies gather when they inspect power lines. So for example, they fly helicopters and they take pictures of the grid. What we do is…instead…this generates a huge amount of data in big utilities. Like Hafslund here in Oslo they collect hundreds of thousands of images every year. So that's one example of, kind of that biggish data is nothing – it’s Huge Data. There's thousands images that need to be processed. So they have people that day in and day out of look at this pictures to find problems and folds. So what we do; we try to apply AI to this to automate as this process as mutch as possible. So we use of the latest technologies that we have today of image recognition and this kind of thing. So instead of recognizing faces, we recognize insulators and and recognize problems.

TM: When you say you apply AI is it like you give up your own kind of algorithms in an AI technology or do you reuse others from other big platform companies? Or in both ways?

DR: We stand on the shoulders of giants. And web use of course everything that is available, and now there is, you know, it is a huge speed in development. Basically every week we see some new exciting development and see opportunities of using it in our domain. But of course all of…all of what we can find as open source or things that are published needs to be adopted and improved to be able to have an optimal application in our in our domain but one of the biggest advantages – well one big factor, is not necessarily, you know, the AI technology or the model that you're using, it is the data witch is one of the most important things. Recently we were in a competition in the US with different companies such as IBM or in Thailand some big Department of Defense contractors in the US doing exactly this recognizing folds on images and we came out on top of that competition. So that we are very proud of, yes.

TM: Ah!

DR: But the reason for that was not much that we used the methods in definitely different from others, but because we had a longer history. We had more data to build our models on so we have been working on this thing with the Norwegian utilities for three years now. And we're able to combine the data from many different customers into our models and that is a huge advantage. So it's not only the technique the machine learning algorithm, but it’s also a good quality data.

TM: Good quality data sets and from different sources. Combines and then you get like…

DR: Yes. Good quality dat set that trumps any good algorithm.

TM: hehe, exactly! So is it is it fair to say that…in regards of the…you mentioned that we need a lot of new investments in the grid because the grid is getting old. We haven't put enough money into maintenance - for actually a few decades.

DR: Hmm.

TM: But isn't it fair to say that your field of expertise could actually make it a little bit more easier…if you succeed my tariffs at home, all the others equals won't be that much higher because you can optimize the load even in a millisecond. Is’nt that true?

DR: Exactly. That's one of the biggest application areas that you have optimizing, you know, investments because nowadays if the energy or the grid operator operates blind, the grid need a kind of dimension for their picks that the extract…

TM: For the maximum pics.

DR: … and usually it's over dimension. Now they have mutch more data so they have an opportunity to analyze this data and find out exactly in which areas they need to invest. We have some customers that found out that some transformer stations were underused while another transformer station was overused, and they were able to swap them instead of buying a new one and replacing the transformer. So there are huge opportunities for optimization of the grid and saving investment costs.

TM : So why we use some time to discuss this area in particular, is that…I think this is one of the areas where AI and Big Data show us that the future possibilities in other kind of sectors. I think that in some way in many cases AI and Big Data is gotten kind of a buzzword. Not that really mature to use, but do you agree or do you think I'm too pessimistic?

DR: Yeah, both Big Data and AI are buzzwords, but you know, now data is really big, it’s not like it was 10 years ago. Today we are talking about, you know, zettabytes of data. I like to take the example of aircraft engines, you know, if you take one aircraft engine it generates 20 terabytes of data per hour.

TM: And 1 terabyte for…

DR: And then you ask yourself if you think about all the planes that are in the air every day and then you multiply for every day in the year, here you come up to like 2,5 zettabytes of data. But also it's difficult to understand, I mean, what is a kilobyte or this zettabyte – how mutch data is it.

TM: Web understand its quite alot.

DR: Yeah, just to give you an idea. Just to take one thing that I used to explaine to people how big a zettabyte is, i use this analogy that you say; okay, if 1 bite is one rice corn then how big would a zettabyte be? How much rice would that be? Okay, and if you do some calculations then once that 1 zettabyte is basically that you are able to cover The Pacific Ocean with rice 4 times over.

TM: Oh, really!

DR: So that gives you kind of an idea what a zettabyte is.

TM: Fantastic! Just to sum up your analogy from from every aircraft to engine in the world. What do we use that data for? Obviously in the airplane, but to optimize or to do any other interactions in the engine industry.

DR: Today we are still mostly gathering Big Data. I've seen statistics saying that we basically analyze only half a percent of the data that web collect.

TM: So your point being that we accumulate a lot of data rapidly faster than we used to do, but web do not use the data

DR: That depends. If you look at companies like Google they're using the data a lot.

TM: Exactly.

DR: And Facebook and all this. But other industries like oil and gas Industries or more traditional industry they're not really exploiting the full potential of all the data that they're gathering, and that's where our role is. I mean, trying to help this company in having an optimal way of dealing with all the data that they gather and achieve some business advantage out of that.

TM: Okay, Davide. It was really nice talking to you.

DR: Yeah, my pleasure.

TM: Good luck with both the american kind of defense industry, was it?

DR: Yeah.

TM: And contract and other major or smaller clients projects. So good luck. And thank you to all our listeners, and for a great job, Davide!

Du har lyttet til en podcast fra Lørn.Tech - en læringsdugnad om teknologi og samfunn. Følg oss i sosiale medier og på våre nettsider Lørn.tech.

Hva gjør dere på jobben?

Vi utvikler primært sky-basert og AI-basert programvare for nett- og energiselskaper.

Hva er greia med big data?

Big data handler stort sett om mengder data som er for store eller er generert med for høy hastighet til å bli prosessert eller lagret på en «vanlig» datamaskin. Big data krever vanligvis sky-løsninger og parallell bruk av mange såkalte «virtuelle maskiner».

Hvorfor er det spennende?

Big data er spennende fordi det åpner for utvikling og bruk av «intelligente» tjenester, som for eksempel Google-søk og automatisk oversettelse fra et språk til et annet.

Hvorfor er det skummelt?

Det kan fort bli overveldende og det kan åpne opp for «big brother » lignende applikasjoner, som for eksempel en stor-skala overvåkning (som er under utrulling i Kina) og andre uetiske applikasjoner

Ditt beste eksempel på big data?

Det må være Google-søk, 40 000 søk i sekundet og 0.2 sekunders svar tid.

Dine andre favoritteksempler på big data, internasjonalt og nasjonalt?

Flymotorer. En flymotor genererer 20TB av data hver time. Hvis vi ganger det på antall flymotorene på et fly, antall fly i lufta og antall timer i lufta, så kommer vi frem til 2.5 zettabytes per år generert av flymotorer alene.

Hvordan funker det egentlig?

Man trenger høy båndbredde og tusenvis av datamaskiner som jobber sammen på en koordinert måte.

Hva gjør vi unikt godt i Norge av dette?

Automatisering av inspeksjoner basert på intelligent bildeanalyse.

Et favoritt big data sitat?

There are a lot more people who know how to move big data around than know what to do with it.

Davide Roverso
Chief Analytics Officer
CASE ID: C0051
DATE : 181012
DURATION : 17 min
Superintelligence av Nicklas Boström
Big data AI «intelligente maskiner»
"Big data er spennende fordi det åpner for utvikling og bruk av «intelligente» tjenester, som for eksempel Google-søk og automatisk oversettelse fra et språk til et annet."
More Cases in topic of AI AND BIG DATA TECHNOLOGY
Digitale tvillinger

Michael Link



Big Data og geometrisk modellering

Heidi Dahl



Slik kan Big Data predikere fremtiden

Sverre Kjenne