The Humanity Behind Smart Animal Breeding

The Humanity Behind Smart Animal Breeding
January 2016 RIDT

Through the compilation of biological data, scientists and farmers are working together to ensure the health of cattle and pigs, increase the quality of the products derived from them, and to maximise their productivity. Here, Dr George Azzopardi explains how this is being done, and how this system could revolutionise the world as we know it.

There is no denying that farming was the singular most important advancement in the history of the human race. After all, it was the knowledge of the cycle of the seasons and the understanding of how crops grow that first led us to shed our nomadic tendencies and settle down.

The rest, as they say and is so apt in this context, is history.

Yet for the world-changing revolution it spawned, farming remained relatively unchanged for millennia and, apart from a few tricks of the trade picked up by the many generation of farmers that ensued, it was the industrial revolution that truly transformed farming from a manual labour to a machine-dominated world.

Today, technology also plays an important part in the growing of our crops, the rearing of livestock, and the primary (meat, milk) and secondary (leather, animal fat) products that they give us.Yet while all this may be one step further away from Mother Nature, the future has never looked brighter for farmers who live off the land, and the animals those farmers look after.

“The idea behind Smart Animal Breeding with Advanced Machine Learning Techniques is to analyse animals’ biological (genetic markers) and behavioural (e.g. quantity of food per day) data, as well as environmental (e.g. temperature, humidity) type of data, in order to automatically determine certain factors that lead to various circumstances,” explains Dr George Azzopardi, a lecturer at the Department of Intelligent Computer Systems within the ICT Faculty of the University of Malta, and a co-supervisor of a PhD student at the University of Groningen in the Netherlands, who is studying Smart Animal Breeding.

“This is done to understand various outcomes, such as what is the best combination of genetics, behaviour and environment that makes a very healthy and productive cow. For the time being, the project is mainly based in the Netherlands, where the dairy industry is particularly important and where farms are already running very advanced systems,” he continues. “These farms are equipped with many sensors that can measure the daily activities of every cow. These include the quality of the milk (by measuring the quantity of proteins and fat, among other things), the number of steps a cow makes every day, how much it drinks and eats, how long it spends chewing, and how long it sits for, for example.”

By understanding the numbers within a context of numerous healthy cows and pigs, in the future, farmers, scientists and veterinarians will be able to tell whether the cow or pig in question is healthy simply through these sensorial observations.

“This modelling technique will also be able to give us early signs of disease and make it easier to treat illness within cattle. Therefore, it will bring the risk of having diseases spreading across a farm, which may lead to devastating results, to a minimum,” he continues. “Of course, this will prove to be vital technology for farms that have thousands of livestock.”

Although the human brain is an enviable intelligent device, it’s not trivial for a human being to determine complicated interactions between many factors. This is where machine learning (a field within Artificial Intelligence) can contribute to applications where a lot of data is available. Machine learning is a term referring to the development of algorithms that are programmed in such a way so as to automatically learn the relationships between the involved components of some given data.

The three farms involved in this project, in fact, have been collecting and storing tonnes of data for the last three years, and there is now enough data to start making sense out of it. The project in which Dr Azzopardi is involved will be investigating and developing machine learning techniques to determine important information from this data.

What is interesting to point out is that this project was initiated by the farmers themselves, who formed a shared consortium with the Dutch government and invested a lot of money in it. In fact, Dr Azzopardi and his colleagues in the Netherlands have received a research grant of approximately €500,000 for this four-year project, which will start in January 2016.

“Yet this project has a lot more potential,” adds Dr Azzopardi. “While we are currently focusing on the animal farming industry, the technology that we will be using is also applicable to other industries, including engineering and healthcare.
In Malta, for example, there are around 40,000 people who suffer from diabetes and who are at risk of developing diabetic retinopathy [when damage occurs to the retina due to diabetes]. Each year, each of those 40,000 people has to have a photo of their retina taken, which amounts to 80,000 pictures that need to be checked manually by professionals.

“Using the same principle, we could teach a machine how to distinguish between a healthy and unhealthy retina, and to simply flag up any pictures that require the attention of a specialist. This will let professionals focus on important cases, and on the treatment of the problem… And this isn’t a farfetched dream, either, as it’s already being implemented abroad,” Dr Azzopardi explains.

Among other things, Dr Azzopardi encourages more Maltese industries to come in contact with the research being carried out in Malta, while he states that “investing in intelligent systems can help maximise performance.”

You can be part of this fascinating world of research, too, by helping many others achieve their breakthroughs in all the faculties of the University of Malta. Please click here for more information on how to donate to research of this kind through the Research Trust (RIDT).



Leave a reply

Your email address will not be published. Required fields are marked *