Velkommen til Lørn.tech -en læringsdugnad om teknologi og samfunn med Silvija Seres, og venner.

Silvija Seres: Hello and welcome to a Lørn conversation. My name is Silvija Seres and my guests today are Emek Seyrek, who is a CTO at Fishency Innovation, and Tomas Finnøy, who is the manager of Saite Consultant and the consultant of AquaCloud. I don't know if I said that correctly, Tomas.

Tomas Finnøy: Yeah, that's correct. Close enough! Thank you.

Silvija: So we'll be talking about sensors and data related to AquaCloud and how do we actually create something really advanced out of fish data. Before we start our conversation, I'm just going to say a very quick introduction to the series and of this conversation in that series. Lørn is doing a series of conversations about AquaCloud which is the data platform of NCE Seafood Innovation. NCE is one of our national centers of expertise related to seafood or fish farming, and it has a number of privately-held organizations, public organizations, and research organizations in its group. One of the goals that it has or one of the projects that it's running now has to do with gathering and then using a large data set or set of data sets related to fish production. The title of our conversation is AquaCloud Sensorics and Standardization and we are going to focus on what is sensor data, what are the main challenges of the industry that you can help with sensor data, why is it important to standardize, how can you help other companies that are related to the cluster contribute to this standardization, and how do we improve the whole industry as a result of this. So, we'll get started if that's okay, and my first question is always who are you and what has made you so. Tomas, in our warm-up normally we always start with the cluster, but actually, for this one, I'd like to start with ladies first, please.

Tomas: Of course.

Emek Seyrek: Yes, so I'm Emek and I'm the CTO of Fishency Innovation, a company based in Stavanger, Norway. We are developing a fish welfare monitoring platform for the aquaculture industry for Salmon fish farming. I'm the co-founder of the company. I'm a chemical engineer by title, a Ph.D. in chemistry, and I have various experiences in Academia and Industry for the development of various kinds of technological products. I've been in fishing seasons 2017 since the beginning actually.

Silvija: I can't guess from your name, where are you from?

Emek: I'm from Turkey originally, yes, it's been a long time I haven't been living there but I'm from Turkey. 

Silvija: Cool. So, you're a chemist by trade, I guess there is quite a lot of chemistry going on behind the data, or is it more understanding basically processes and modeling those processes? I think it's fascinating to see how our education forms us, so, I'm just curious.

Emek: Well, my chemistry background has nothing to do with what we are doing right now. I'm more applying my engineering skills to coding and in the development of new technologies. Of course, chemistry is helping, chemistry opens up a lot of windows to everything, so there is some kind of chemistry there. I did a lot of computational chemistry and biotechnological development, so that is how it is related. But chemistry is everywhere. We cannot get away from chemistry, including the fish. There is a lot of biochemistry and biology, more than chemistry, that are big challenges for the industry, for us, for everyone.

Silvija: Yes, I won't spend much more time on this learning about science, but I believe that as much as this last decade we have focused purely on computation and it's a binary mathematical logical computation, the next 10 years are going to be chemical computation. So, to me, it sounds like your background is where we are going because modeling is increasingly the physical world, which is following some other logic than the basic mathematical logic that we have in our computers, so it'll be really, really interesting to see where it goes.

Emek: Yes, I don't want to go too much into that either, but yes, chemistry, physics, biology, and computation are meeting a lot everywhere right now, and Mathematics, of course. That's what we are doing altogether.

Silvija: Very good. And Thomas, who are you?

Tomas: Yes, I'm Thomas Finnøy. My background is not from chemistry but from electronics and software development, I took a university course in Newcastle almost 20 ago and I worked as a developer in Norway for different branches: first was the military, where I was developing software to integrate systems from different suppliers, very similar to the problems we are now trying to solve for the aquaculture industry. So, I basically have this issue in many different areas of business and the solution always ends up being the same. I've been in aquaculture industry related work for seven years, and also then AquaCloud the last two years, working with almost the same, integrating systems and how people work with systems, making sure that data flows as it should, and give them the right information at the right time to the different actors in the companies. AquaCloud is what I'm focusing on here now, the sensor data standard project was something me and the other managers started working with, and they agreed that we should try to push this idea into the industry right away because everyone needs this type of solution.

Silvija: Very cool. So, tell us a little bit more about how does AquaCloud work with sensors?

Tomas: Yeah, so the AquaCloud sensor data standard is a project made to make a standard for how systems should communicate sensor data. That means that when a supplier wants to develop a system that is using sensor data in some way, it should develop it around the standard using the definitions that the standard defines. So, its system internally and the other system communicates does it in the same way. 

Silvija: Can I interrupt? I'm trying to make this a little less buzzworthy. So, I'm a fish farmer and I want to join your platform. I'd like to be data smart and I don't have to reinvent the wheel, so I don't need to start figuring out myself which sensors I should buy, how I should put them up, connect them, and what I should do with them, all of that is described somewhere in your standards, right?

Tomas: Exactly. I like to talk about an analogy sometimes to make it simpler for people to understand what we're doing. Basically, if you think about a house and the electricity system in the house, what we're trying to do is making sure everyone uses the same type of plugs in the house and the same type of voltage. So, if you buy a lamp from the shop, you know that it works in your house and you don't have to talk to the seller asking what type of version the lamp is, who made it, when was it made in, and has it been upgraded lately. So, whenever you buy something from any supplier, it should work with anything, and not only lamps, any extension cable, or any type of device you have should work with any other type of device that you want to use in your house. That's what we're trying to achieve, a common way for everything to interact in a simple way.

Silvija: So you're helping them provide data pretty much that you can use, that you can cross-connect, because you know then that it will be not just from the appropriate kind of lamps, but that the lamps are placed in the right depth or height, that there is a set of principles behind data-gathering because I assume that, if not, then the job that Emek has is completely impossible.

Emek: Just to make easier to understand the overall picture from the supplier side, we don't have to go to each farm or each user for our unit asking what do they want, because there is a standard already which helps us a lot to know. This is what we need to do. This is the way we need to provide the data, so it shouldn't be confused on what kind of sensors you should use, etc. It's about how the data should be, how should it be communicated. For the supplier, for us, that's the most important thing because customized service is much difficult than standardized service as Tomas points out. That's the main goal for us.

Tomas: Basically, there's a lot of data out there in the industry, a lot of devices having gathered this data and storing it in different places, but issues for suppliers like Emek is that it takes them a lot of time to figure out about what data is there, how it's stored, and also, how to use it, maybe the data quality is unknown and such things like that. The things change over time, too, so this farm might have used one type of sensors for a while, but they wanted to change them because maybe they don't work anymore, and then, I would have to know about this and customize the system to work without new sensors as well. That is very complicated and involves so many people. It often fails because things change, people don't really know that it has changed basically because not everyone is aware that they have to tell everyone about it. We're trying to make sure that whenever someone buys something new it works with everything else by design.

Emek: I like your analogy very much Tomas about the electricity at home. Everybody cannot use different plugs, it's just not feasible, not sustainable. Sustainability is what we talk about all the time nowadays and this is very important.

Silvija: Long-term integration and long-term perspectives. Help me understand, what kind of data are we talking about? I can imagine measuring the temperature of the sea and maybe I can somehow imagine using AI to count the number of fish, but it's all a blur. So, what data can you gather.

Tomas: Our first version of the standard we try to make it simple and easy for us to communicate. So, what we focus on first was the environmental data measurements like temperature or sea-line levels, and then oxygen levels and such. Also, we wanted to add something about feeding. We have to have feeding data types, which are basically the current feeding temple and the amount, and also the feeding silos amount at any time, so you know how much you fed and how much you have left in the localization. I would like to add that we are very interested in developing the standard further, so anyone that needs certain data type into their systems, we would want to talk to them to see how we can facilitate that and develop the standard further. We are in talks now to join the project on sonars, which you can do a lot with, for example, measure the amount of fish you have, or where it moves in the cage and such. Also, we are in talks about extending it with better light measurement, so you can have multispectral light measurement indicators as well. We are reliant on the industry and suppliers like Emek care to tell us what they need and if something is missing in the standard, of course, to contact us and tell us about it, so we can run a process on it to upgrade it because this isn't something that stands still, the industry will change faster and faster the better standards we have.

Silvija: Before we throw the ball to Emek I have to ask you a question. We had the previous conversation in the series with Björgólfur and he talked about the case from Canada where they used also satellite data and imaging to track LG development and movement. So, not all the sensors have to be within a particular fish farm, there could be also external sensors if I understand.

Tomas: We don't have to look at two different projects in AquaCloud because it's formed by several different projects. So, what our project is focused on is what is measured locally in the cages and the sea around the cages being directly connected to some former system. Any suppliers can combine these data with such things, like your weather data from satellites and stuff like that, of course. There have been talks about adding microbiology sensors in our project which could be combined with this type of system like you are mentioning which could feed data to this type of overall system and then this price can start again. Our standard is focused on gathering the data locally.

Silvija: So, Emek, what do you focus on?

Emek: We are actually more looking for the advancement in another part of the AquaCloud project, which is the fish health, which is as Tomas said, there's the sensor data standardization and fish health standardization. So, the goal is, as Tomas said, multiple data needs to be gathered. Since we are a fish welfare monitoring, we have a fish batter monitoring device and fish health parameters, and there's a lot of very important information that the farmers need to know, that they need to communicate with each other. There are also public data that they need to share between farms and within the farm itself between different departments. So, of course, we also have other types of sensors in our device as well like a temperature one for example. What has been developed already is helping us a lot but we are looking for further development of this standardization project to come up with standards for other aspects as well, such as fish health.

Silvija: I know nothing about how fish health is really measured. I have a couple of point examples from startups that I admire, so, the picture I have in my mind is that you have some cameras and they can count the fish, they can see how much the fish has grown, they can maybe see patterns in fish movement, maybe they can look at fish skin. How do you recognize the happy fish?

Emek: Yes. Happy fish is very important and one of the most important problems is the sea lice, you must have maybe already talked about it. It's a parasite that is found on the salmon, it's one of the biggest problems of the industry, and it's on the skin, as you say. So, basically, we look at all the visual aspects with a camera system that monitors the fish and look at what kind of abnormalities or health issues can be present on the fish. We have a 360-degree view of the whole fish, so we are able to see all sides of the fish as if you were in a lab, but you are not, because the most difficult part for the farmers is that it's not like a pork farm, you don't see and you just catch your pork and look what is going on. It is very difficult to monitor underwater fish and far away, but in the end, it still has to be done. There are also regulations and they have to do some kind of manual analysis, they go out in the sea and pick up 20 fish out of 200,000 and they try to make an assessment. What we are trying to do as a solution is to see everything underwater, so you don't have to touch your fish and your fish can stay happy because you have to take them out for analysis and you also anesthetize them, so it's not a fun task neither for the fish nor the fish farm. It is to help this process of fish monitoring for the sea lice and for other health issues that can be visually visible. But it's not only for the skin, you can also see how stressed the fish is, the way they swim, the way they move, and you can't do it when you just take out a few samples from the scene.

Tomas: There's also several levels of standardization here, so, what Emek is talking about is how you can measure or see how many lice the fish has and you have different solutions to do that. You can have cameras or multispectral camera, so you can, of course, do it manually, and even lasers today, that let you see the lice and shoot them, and on the low-level gathering that type of data in the large databases, is very interesting to see how good the systems are working and how well, for example, a treatment works against the number of lice. But then also, you have, of course, the standardization on top, which is saying: what is a healthy fish? How many lice can we accept on a fish? And how does it affect the fish's happiness over time? All these things we need to agree on how to define them so everyone is working on the same platform, and systems like Emek's and others are the foundation for gathering that data and analyzing it later.

Silvija: Again, sorry, forgive me for having to ask my dummy questions about fish farming. The technology that you're developing is specialized for salmon, but could it work in theory for other kinds of fish, and even maybe for other kinds of seafood or crustaceans?

Emek: Absolutely. Basically, of course, it will need further development for other species but it is applicable for other species and other kinds of applications. Technology's where the world is going anyways and all these kinds of AI technologies are supposed to answer all these needs for different applications.

Silvija: So, my question to both of you has to do then with developing the industry, and please correct me here if I'm wrong, but my impression is that Norway is actually quite advanced in both its technology and infrastructure for fish farming but also in its regulation and it is one of the reasons why we're actually able to sustain that industry. In some other countries they have gotten for example the lice problem completely out of control, so there is some sort of export of ideas potential here as well.

Tomas: Yes, we are looking at working with the industry outside of Norway. We have been to talks with interest groups in Ireland, they've talked about their situation now, and it sounds very much like it was in Norway over 10 years ago, which is where the sea farmers themselves didn't have much understanding or contact with their data sources or the data stores, and they were very reliant on the competence of the suppliers to develop solutions for them, which could be a very complicated matter when the supplier doesn't really understand the need of the sea farmers at any point in time, the communications get difficult to get up and running. So, having systems and standards like this moves the sea farmer closer to the data sources and they can require more detailed specifications, which again creates more and better-detailed systems useful for the sea farmer. Of course, the standard as it is now is in its infancy and we would want to push this into international organizations like ISO to make sure that it gets a foothold there, and, of course, on sea farmers that have international connections, like Chile, and Canada and everything. So, we are focused on trying to spread this as much as possible to make sure that we don't have competing standards in the market, which would be difficult to handle for the suppliers. There is a little bit of a technical point, but our standard is based on what's called OPC UA standards, which is a massive corporation, which is doing standards for sensors and automation, and that's also a way we will push this specification into, to make sure that anyone that is using OPC UA in aquaculture will be able to use our standard for that.

Emek: I would like to also add something about Norway. We are certainly very lucky that Norway is where it is at a technological level and that they are willing to apply that to our aquaculture because in Europe we are a bit advanced, but if you look into the details of the technologies that are involved in making aquaculture work better and be more sustainable, I find it quite amazing, and it's gonna definitely come to other countries, I mean, it's the future for fish farming and everything else. So, Norway is definitely ahead, very much ahead in this.

Silvija: I think that's an important point to make, and it has to do with the digitalization of our industries, but also our future. This question is perhaps directed first to Emek as you are a foreigner like me. I think Norwegians really underestimated the level of sophistication of infrastructure in this country. I mean, we still complain about not having a bandwidth of 5G technology in some remote corners of this country, but compared to the rest of the world, maybe with the exception of Singapore, we really are incredibly networked and data-rich.

Emek: Yes, definitely.

Tomas: It's not often that I walk around with my phone up in the air trying to find the signal in Norway, but I'm quite used to it when I'm traveling on vacation, so, I think it's very true.

Silvija: But also the fish farms. These things are way out there expecting 5G, Internet of Things, and all kinds of other fancy things like AI-enabled technologies, that you are doing Emek.

Emek: Yes. For example, in the sea farms out in the sea we might not have good 4G, there are always other solutions and it's not easy to be connected somehow and apply technology like that, and to access data is not that easy in many other places, but it is possible in Norway because the infrastructure exists, so you're absolutely right. Even for the 4G, the companies' service providers and the network providers are working to provide solutions, they know that this is important and this is where everybody will be needing. There are projects to make even 5G in some regions, I mean, they are already started with 5G for these kinds of applications.

Silvija: So, if I asked you for your main kind of three requests for the listeners here, perhaps both within and outside the industry, who should be interested in learning more about your standard, and perhaps, participating in it? What should they do? Maybe we start with Tomas again.

Tomas: I think almost anyone that is interested in getting better data, better reports, and better communication within their company and with the cooperation with other companies, anyone that's working in purchasing or developing any type of internal systems reports would want to be interested in why they need to standardize, and, of course, suppliers that want to grow and not be hindered by complicated integration issues and complexities, and want to work with other companies seamlessly and efficiently, want to be valued instead of someone that is hindering value-adding and systems development. It's not scary, it doesn't block you from doing anything, it can be an addition to what you already have, or it can be just something you use for a certain set of customers, not everyone is that neat. So, basically, just engage with us and try to understand if there's something that is not clear, and we will help you as much as we can to get you up and running. So, please, contact us.

Silvija: Emek, I'd like you to add some international perspectives, as well. Let's say that I'm an Oil Engineer. I think that there is quite a lot of transfer of knowledge opportunity here to all related industries, like agriculture, or oil, or infrastructure at sea. What do you think?

Emek: Yes, the goal of this is actually standardization. If you look at the hierarchy of the information that you get, it starts from the country, for example. This one, of course, it's specific for aquaculture, but it will be possible to include another hierarchical level to make it applicable to other industries. So, this is why it's very nice to have such work already, which is easier for implementation in different directions in the future. And, as I said, in the standard there is the country, so, it starts with Norway, but it's not specific to Norway, it is worldwide applicable already.

Tomas: As you said, since we're using this OPC UA platform, which also Equinor is using on their platform, any other industry that is gathering data can use the same type of platform, where you can connect all this together in a large data platform reporting system and the suppliers can basically create a system that works for any industry. We were in talks with a smart sea project, I can't remember what company was behind it, but it was a large scales data-gathering project using all types of industries data for reuse by anyone. We, of course, aquaculture, the oil industry, shipping, and everyone was going to be involved, and, as long as everyone is using UA to define their standards, then anyone can communicate with any system that is built upon, and it can be used in all these branches of industry. So, absolutely, this is relevant for any type of data gathering industry. 

Emek: And for us, for example, we are not only talking with fish farmers, there are fish health companies and also fishing companies, because this technology can be adapted in a certain way by modifications, or depending on the problem you can definitely switch to different industries, and having a standard like that is definitely helping a lot.

Silvija: I imagine even environmental companies of different kinds in open spaces. I've just looked back to my notes from the first conversation I had in the series and, well, currently, we are producing 1.1 million, tons of fish annually, and the goal is to move to 5 million tons within the year 2035 or 2050 perhaps. So, I would like us to basically, conclusion-wise, play a little bit with this number because if we are actually going to quadruple or quintuple the current production levels, we are absolutely dependent on doing this in a sustainable way. So, without your system, is there any way to even think of a goal like this?

Tomas: I can't see how this could be done without making everything much more efficient and much more sustainable. Of course, making sure that the environmental footprint is as small as possible and having under control that there are no very big surprises suddenly coming as we had in Northern Norway last year with the blossoming of the algae, we can't be at least reason for that to happen anywhere else, for example, so we have to control what we're doing. I cannot see how we can do this without standardizations and with systems like Emek is making, where we have full control over how the fish's life is, basically.

Emek: Yes, and, of course, with such a scale-up in the industry, you have to scale up the way you process the stuff, so, you can not do it one by one. There needs to be some kind of way to standardize and this is the way to go.

Silvija: Very cool. Emek, by the way, you gave us some really good reading tips and I'm just going to read them up super quickly, it's The Rise and Fall of the Third Chimpanzee by Jared Diamond. I haven't read it, do you think I should?

Emek: If you're interested in how things have changed, it's quite an old book, by the way, but it's quite amazing how things develop and change. I think it was written in 1993, and when you read it again today, you see the change.

Silvija: Yes, but you also recommend the Queen's Gambit and the Fishency 360 underwater salmon videos.

Emek: They are my favorite videos nowadays! Basically, it's how we do our analysis, the fish swims through the unit, and then, we look at the fish, of course, we do it automatically by AI, but to do some kind of quality controls is my favorite to watch in, not really in free time, but it's quite fun, and to see the fishes it's relaxing, like watching an aquarium. Sometimes I leave it on to watch a few stuff and my kids before sleep kind of have a look at this.

Tomas: I guess I can suggest the people reading the sensor data standard, maybe it will keep people awake but at least it's interesting.

Emek: Yes, it's a good read!

Silvija: Tomas Finnøy and Emek Seyrek, thank you so much for joining us for an inspirational and learn-rich conversation.

Tomas: Thank you very much.

Emek: Thank you.

Du har nå lyttet til en podkast fra Lørn.tech- en læringsdugnad om teknologi og samfunn. Nå kan du også få et læringssertifikat for å lytte til denne podkasten på vårt online universitet Lørn.university.