Machine Learning In Today’s Society
Machine learning is a very big part of society today and it is gaining a lot of interest recently. The way machine learning works is like how humans learn. So it helps to understand how humans learn to understand machine learning.
So, humans learn by building models. We look at our environment. We collect data through various inputs such as our eyes, our ears, our skin. As we walk around the environment we are collecting data like pixels from our eyes. Then we build models around what things are from the data we are collecting.
Those models help us to interpret some of the things we see later. So, if we see a picture of a cat, then the next time we see a cat, the brain knows it is a cat. This is because we build models around how cats look in our brains.
How Machine Learning Works
What machine learning involves is building models, but this time around it is not humans. It is building models from data. Now, we take data and we put it inside an algorithm and that algorithm spits out a model. In the past we usually program everything. So, if we want to identify a dog we say it has four legs, it has a head, it has eyes. But machine learning says no, don’t do all that. Take as many pictures of dogs as you can find and tell us that that is a dog. You label that data. You take a lot of pictures of dogs. You say this is a dog, and this is a dog. Or you take pictures of cats as well. You say this is a cat and you put it inside an algorithm.
The algorithm looks at all the dogs and looks at all the cats and on its own builds a model of how dogs and cats look. Then it spits out the model. Then let’s say in the future you put in another picture that it hasn’t seen before. It will take that picture, run it through the model it built and say if it is a dog or a cat.
The Different Types of Machine Learning
There are other forms of machine learning like unsupervised learning and semi-supervised learning. Also, there is reinforcement learning. But it all boils around building models from data. Machine learning is important because humans are terrible at making good decisions. We see ourselves as very intelligent. Yeah, we are very intelligent compared to other animals. But if we look at the kind of decisions we make, oftentimes we make biased decisions. That is a function of the data we have.
Every human collects data. From that data, they draw conclusions or make decisions. But if that data is faulty then we have a system where almost everyone makes wrong decisions. In the case of machine learning, we can have a machine collect data from all humans and as many humans as possible. So the machine collects as much data as possible from the wisdom of the crowd. Then with the right set of algorithms, it will get rid of the bias. That is if we say to get rid of the bias in that data set.
Smarter Than The Human Brain
So, there is a high likelihood for the machines to make mistakes or wrong decisions due to bias. It can also be due to the unavailability of data. Humans have small data at their disposal but can have a machine that has access to a lot of data. Also, we live in a big data era with computational power going up. This means we can crunch thousands of numbers. Numbers that a human brain will not be able to crunch.
The machine learning will crunch these numbers and help us make the right decisions. An example of this is weather forecasting. I doubt there is any human brain that can keep track of all the weather that has existed. Human brains can’t make the right predictions of what the weather tomorrow is going to look like.
The machine learning has been successful in the folding of proteins. There are so many applications now, such as self-driving cars. We are looking at building vehicles that drive themselves. This is because humans are terrible at driving. So, it’s important.
What Machine Learning Is Up To These Days
There are military applications and drug discovery. There is so much data that humans haven’t been able to process and pull insights from that. With the help of machine learning, we can see things from those. We can process data that we have been accumulating since the beginning. So, it is important because it is like having gold sitting right under your basement and you don’t know it is gold.
We are sitting on silos of very important data but we haven’t been able to draw any insight from them. But with machine learning, a lot of companies today are drawing business insights. Scientific discoveries are being made from all that data that they have been sitting on.
Machine learning can do everything a human can do and a lot more. So, as I said earlier, machines can crunch large chunk of data and pull insight from them. An insight that humans would have not been able to do.
Finding Cures And Developing New Medications
There are a lot of applications now in drug discovery. This is because there are all these chemical compounds that we want to find out about. If I take this chemical compound and I mix it up with this chemical compound, what will be the result? Will that lead to a drug that we can use to fight some diseases?
Machines can crunch thousands and even millions of all these compounds. They can build models around them and create simulations that we can use to discover drugs. Humans, yeah, one brain can do that if you sit in your lab and mix all those things. But it could take you thousands or millions of years to do.
Creating Innovative Robots
In logistics as well, we can bring down overhead if we apply a lot of machine learning in warehousing. For example in the robotics industry or using robots to pick shipments. Efficiency goes up if we apply machine learning. For example, deciding on routing for self-driving trucks. This includes optimizing the route so that we save on fuel. If you have a self-driving truck you could decrease the number of accidents from the road.
The military is also looking at facial recognition. They want to use drones that will run themselves. The drones pick a target and get rid of the target without any civilian casualties. The list is endless.
It applies to anything you can think of. For example, the printing of houses. We can apply machine learning to 3-D printing and 3-D print houses and have robots do everything. In agriculture, we can have automated farming from scratch. This includes planting, harvesting, and distribution. This could get done completely without humans with the aid of AI-infused robots.
Machine Learning Keeps You Safe
Many companies are building AI security agents, for example. They scan the environment using facial recognition software to identify people. They run those faces through the police database to find out if any of the people around are a threat. So, the list is endless.
As far as in the stock market we have a lot of robots now taking historical data of stocks. They are predicting what the stock price tomorrow is going to be.
Natural language processing has grown these days as well. We have agents that make it very difficult for you to distinguish if the voice you are listening to is a human or a robot.
How Machine Learning Has Grown
So, the desk that used to be the benchmark for AI has changed recently with Google and many other companies. The list is endless as far as what we can do. This is in our current confines more or less. We aren’t talking about general artificial intelligence. Now they have systems that learn on their own and build models on their own. We are talking about this current supervised machine learning. Sometimes unsupervised machine learning is apart of that too.
Machine learning has no limits. All we have to do is have people think about how to solve problems that we couldn’t solve before. Pick the data and build the right models. Machine learning will help us do things that the human brain never could.