Artificial Intelligence is the field where we try to build systems that perform mental tasks just as we would ourselves, perhaps even in a better way. These could be simple tasks like finding a name in a list of guests, or something more complex like reasoning about the Nature of Man, and everything in between. But what about robots taking over and governing our lives, you say? Don’t worry about that, as the route to Artificial General Intelligence is still very unclear – ask Yann LeCunn.
So, we need a system whose inputs are some sort of data and its outputs are completed tasks – but where would one even begin? Well, you could start in the 50s with some simple math and human logic: give a computer some data about the problem you are facing, and give it some rules to behave like you would ask a child. Supposedly, if you gave it some good data and some clear rules, it should handle the task as you had expected from the beginning. Congrats, you have invented Symbolic AI and leveled up to Programmer!
However, what are we going to do for complex tasks like detecting if an object is in a photo, or how a tractor can navigate itself through a field? Are we expected to write millions of rules for every single change and unpredictable situation? Nah, that’s what computers are for, let them figure it out by themselves! This is where Machine Learning comes in: instead of telling the computer how to handle your data, we shall give it some previously gathered data and what the expected outcome is (training) so that it can decide by itself what set of rules it needs to follow later on in the field (inferring) where we will give it the latest available data. As you can imagine, ML is based on many ingenious statistics. Do not expect it to be faultless, or even understand your problems fully, but it may be just enough for what you need it to do.
Machine Learning, along with IoT and BigData, can bring many solutions to the problems today’s farmer is facing: yield estimation, autonomous tractor/drone navigation, image detection, knowledge extraction from text, automated harvesting robots, and many more. No more are you bound by the lack of hands or expertise – a computer can intelligently alert you about the condition of your field, act on your behalf, and assist you in your harvest. Exciting times!