[00:00:31] Welcome, everyone. My name is Nikolai Shchetikhin and I have Jim Johnson with me. Hi, Jim, how you doing? I’m doing good, Nick, and you? I’m doing wonderful. Thank you. And today, we’re going to talk about very interesting topic as usual. And today, this will be artificial intelligence and everything related to it. Things like machine learning and reinforcement learning. And we’ll try to answer the question, whether it’s technology helping us in the industry, or it is still more of a marketing hook. 

[00:01:09] It’s interesting subject, Nick, last week, I was reading that by the end of 2024 75% of companies will move from piloting using AI to actually practical applications of it in their everyday business. That‘ll increase the amount of data being used for streaming by five times while it’s happening today. 

[00:01:32] Oh, wow. And I also know that within this pandemic context, we’re already using artificial intelligence to predict the spread of the virus and measure how the effectiveness is going with the countermeasures, right. Yeah, and   

[00:01:54] I mean, there’s other smart ways, smarter ways as well, in terms of reinforcement learning is being applied to really do more complex problems. And in terms of our own logistics, particularly. 

[00:02:06] Right, so let’s look at what this terminology means and what it consists of, because it’s all over the news and all over the media, artificial intelligence, machine learning, reinforcement learning, but what does that mean? So artificial intelligence is a general term. And machine learning is one of the main approaches in artificial intelligence. And under its umbrella, there are several different ways you can do it. And one of them is deep learning is when you use a lot of data, a lot of historical data. And you can provide some insight or system rather giving you insight on what’s going on in it trying to simulate humans brain operation and behavior. The other method is reinforcement learning is when a computer is trying to go through the process of attempts. And through the errors and successes, it’s improving itself, behavior and trying to figure out the optimal solution, more or less like you going through the maze and finding an exit. And there is deep blue reinforcement learning, which is a combination of deep learning and reinforcement learning where we use historical data in large, and we also going through the maze. 

[00:03:40] It’s kind of interesting that explanation, but it does show the importance of your data data is, is the cleanliness of your data, the credibility around your data set is very, very important if you’re using machine learning techniques, or deep reinforcement learning. And it’s critical that because these models are using historic data to train them, that if you’re looking for accuracy of results, that data set needs to be accurate and credible, before you start training your machine learning tool.    

[00:04:15] Yeah, exactly. Because if you think about it, and the result can be quite critical. It could be some  business decision that is at stake. And also, like if we if we think about what we have right now in the real world as an examples of usage of artificial intelligence. So for example, we all know about those self driving cars that are using those technologies already. There are apps that are trying to help you and suggest you how to invest your money. There’s facial recognition technologies and with that, for example, there is quite popular, deep face technology that can put your face on anybody’s head face, for example, in the video, so you can end up in Tarantino movie. And also, there was quite scary example, because people trying to use those technologies and find a ways of using them. So Facebook in 2017, they created to chat bots, they put them on two devices and let them talk to each other. And after a while they developed their own language, nobody could understand what they’re saying. So they had to shut them down.   

[00:05:33] Wow, that’s amazing. And there’s another real application of machine learning within credit card. So American American Express, they process something like a trillion dollars worth of transactions, and about 110 million card holders. So they’re using machine learning just to check for fraud than any other transaction that they go through. And you can imagine the sheer volume of data. So they’re using machine learning in order to try and look at testing where fraud is happening, and sorting that out. But also leveraging the data flows, to really connect cardholders with new products and different service ranges offer currencies and stuff offerings. So as a lot happening with machine learning, and in the real world today.  

[00:06:19] Yeah, many applications and like, like bigger companies like Google is using machine learning to help their internal processes. So they’re use those algorithms to tune their hardware and coolers at their data warehouse to get the better sustainability of their business. And if we talk about more about like agriculture, there is a john deere who’s already using artificial intelligence and visual data to define whether they need to spray crop with a pesticide. 

[00:07:00] Yeah, well, let’s look at it. I mean, in the poultry industry, let’s come back to home here. I mean, there’s a lot of opportunities to use AI within a poultry industry, some of them would be looking at planning, sort of forecasts and in plant and improving accuracy of forecasting and planning, and crunching large data sets to look at troubleshooting, where there are problems in businesses, looking at scheduling and logistics, maybe live birds, getting them from the farms into the into the processing plant and managing and optimizing that logistics challenge. So there’s a lot there. And something we spoke about a couple of weeks ago on a podcast was a IoT. So again, with the IoT, and that’s where we take in information from the broiler house. So the breeder houses in real time, this huge amount of data sets, and that a lot of detail information. So again, applying artificial intelligence tools can help us crunch that data. 

[00:08:01] Yeah, exactly. So for example, like you said, there are like lots of complications in planning or in projections. and machine learning really can help with that. Because, for example, projecting the weight, it’s not only just projecting the birds that are currently sitting on the ground, and will be processed in a couple of weeks. But it’s also taking into consideration all the previous history of previous flocks for that, for example, farm or shed, or that area, or that region. So machine learning can really help here with projecting the weight and improve the accuracy of that projection. And if we’re talking about more complicated scenarios, where you really need to figure out what is the optimal solution for this or for that we can talk about planning of the placements because you need to look at both ends for the supply and demand and figure out what’s the optimal placement schedule for the birds in order to have your eggs produced in order to push them through the hatchery and have the chicks and then have the right number of birds at the proper weight to plan that’s a lot to consider. And this is where reinforcement learning can help because it’s trying to figure out that task, but attempting and error and success. So that’s really exciting. That’s I think that’s the future   

[00:09:39] Yeah. So I mean seriously, there’s a lot of areas where we can apply artificial intelligence in the in the poultry industry today. And there’s many ways it can help us improve the accuracy of planning and forecasting provide insights into production start to be more prescriptive about how we should grow our birds to optimize to become more competitive and ultimately to add profitability to the business.   

[00:10:05] Yeah. And that’s really good. And like what you mentioned that the system is giving you insight or giving you like a prescription. But we still need to have proper people in place in order to take the decision. Because if a system is giving you a result, or giving you a suggestion or giving you a projection, you still need to have that experience in order to take the right decision, where do you need to go? And with more and more technology in place, there still will be a very high need in people with a good experience over there.    

[00:10:50] Yeah, that’s right to emphasize that Nik and you know, just wrapping up really, it’s been an interesting discussion, we’ve talked about artificial intelligence and how we apply machine learning and reinforcement learning techniques and where they can be applied within the poultry industry within the modern poultry industry. And I guess it’s all really about how do we leverage that data set? How do we become more effective in the way we, we analyze and we utilize the data we have in the poultry industry? How do we use the software, the modern software tools to help us improve our performance and become more competitive?  

[00:11:24] Yeah, exactly in Thanks, Jim, for this real nice wrap up. And for those who don’t want to miss any of the new episodes, please join us on LinkedIn, on Instagram and on Twitter. And I will see you in the next episode. Thank you. See you there.