What might a college term paper, a letter to your mom, and a dairy ration formulation have in common?
The answer, when it comes to term papers and the letter to your mom, has been in the news of late. As it turns out, artificial intelligence (A.I.) has been creating content for the everyday person. Why spend time thinking through various approaches and styles for your next blog post, writing assignment, or even a letter to your mom? You can subscribe to a service that is a better writer than you, has perfect grammar, and will assist in completing the task in no time flat.
Is formulating dairy rations next?
Let’s dig in a little and see.
Building a dairy ration has much the same feel as creating something new, just like writing. Think about the delivery of the message in a document like delivering nutrients in a ration. The rules of grammar, style, and form can be likened to nutrient requirements, ingredient choice, and, finally, bunk presentation of the ration.
So, if A.I. can receive thoughts and concepts and churn out a term paper, it seems easy enough to tell a machine about the animal we are feeding, the feeds we have at our disposal, the cost of each feed, and then expect the perfect ration to be created. Yes, all of this is possible and is likely in the process of being crafted if not in use already. How, though, does this new capability hit the farm gate of a dairy?
Part of the A.I. process proposed above for rations is already in use. To be fair, I am out of my area of expertise to really write about machine learning and the details of A.I. However, some aspects of it are already employed in some diet formulation and have been for quite some time.
If you Google the terms A.I., linear programming (LP), and optimization, you will get enough results to indicate that these concepts have some overlap. In a letter to your mom, you might tell the A.I.-powered machine a few keywords about how you feel and your upcoming schedule, or it can just look at your Google calendar. It can access your recent travels based on your mapping app and last but certainly not least, the machine knows your recent shopping habits and read your social media posts. Voila, your letter is ready to send. Your mom will be thrilled with how much you have improved in your grammar and overall writing skills.
If A.I. can write better letters, can it also build better dairy rations?
Replace the online info about your social life, travel schedule, and shopping habits with the weather in Brazil, a recent USDA Cold Storage report, the price and basis for corn and soybeans, the five-day weather forecast, and the estimated attitude of international bullies. By telling the A.I. machine this information and a nearly unlimited list of other facts, we can have the very best ration considering all of these factors.
Wait a minute, though.
Aren’t we already doing some of this?
Yes, but the machine is infinitely smarter than we are. How can we leverage that fact while still remembering that we are going to use a loader bucket to implement this perfect ration into a very imperfect, at best, dairy farm environment?
The possibilities of A.I. in feeding dairy animals should excite us all. Not only can we better meet the nutritional needs of the animals for production, reproduction, and cow health, but we can also ask the machine to plan for this with not just economics in mind but also concepts including sustainability. Now, let’s bring this all back to the dairy for an example that leverages part of what we do now and allows a true A.I. solution make us even better.
Consider a high corn silage diet on a dairy with inline milk components testing, including milk urea nitrogen (MUN) with real-time reporting, an on-farm lab analysis solution for things like starch in corn silage and protein in alfalfa, and constant communication in the cloud with the milk plant to know what milk components were needed that week. We can add to all of this real-time information on ingredient pricing, potential ingredient logistical problems, and the expected weather.
Knowing all of this, we could potentially feed a slightly different ration every day, not to mention to each individual cow if robotics are involved. Looking simply at the starch in corn silage, the current components and MUN, and the protein level in the alfalfa, we could slightly adjust the feed rates of corn silage, corn grain, soybean meal, and soybean hulls to make a better ration in real-time.
Once again, wait a minute.
Do we really need A.I. to use all of this information to build this better ration?
The true answer is no. We could actually do this today if we set up the system to gather it all. We still need a plan that is from the mind of a trained nutritionist to tell the formulation system what to do. But do we really? Here is the point. The A.I. machine can better learn how all of these individual factors truly work together and, in some cases, actually create tomorrow’s reality. This is the part where the machine is truly smarter than us.
Recently, I watched a movie called The Imitation Game. It reminded me that these concepts are not new. During World War II, the British government found the top mathematicians of the day and created a team to break the German code language created by the enigma machine. Alan Turing, known as the father of modern computing, created a device with a lot of moving parts that used logic to break the German code. This discovery had a significant impact on ending World War WII.
While I am sure to be missing many details, this discovery, many years later, offers us the opportunity to use machines that are smarter than us to better feed our cows. So, will I eventually be replaced by a machine not only to formulate dairy rations but also to write articles like this about it all?
Maybe, and that is okay.
What a great opportunity this new level of expertise and work efficiency offers us to spend more time trying to better understand the biology and psychology of dairy cows and the people who care for them.
On to the future!