Sci-Fi is Coming to Life with Artificial Intelligence
We don’t realize this, but Artificial Intelligence is everywhere in our day-to-day lives.
In fact, every time you use a smart device, like a tablet or a computer, you will use AI in a way. The Global AI Market value is at 327.5 billion US dollars and it’s only continuing to grow exponentially.
In the future, they say that AI has the potential to take over the world, which is why it is extremely important to know what it’s about.
This article covers what AI is and how it works, different branches of AI (Machine Learning and NLP), different applications, and where we are with AI right now.
Intro to Artificial Intelligence
Artificial Intelligence (or AI, for short) is exactly as it sounds. It’s essentially creating machines that think like us humans; hence, the word “Artificial Intelligence”!
Today, AI is mostly used to achieve tasks that are usually done by humans, but with a much more efficient way using machines.
In the future, we will see AI advance even more. Currently, we are in the Artificial Narrow Intelligence stage (ANI) where we can do smaller tasks with AI. The second stage is Artificial General Intelligence (AGI) where a machine would be able to comprehend an intellectual task that a human can.
Lastly, we see the final stage being Artificial Super Intelligence (ASI), where AI can complete tasks to an extraordinary level of intelligence, surpassing the brightest of minds of humans.
Currently we are in the Artificial Narrow intelligence stage where we are exploring how AI can be used to replicate ordinary human tasks with machine learning by doing smaller tasks.
There are three main branches of AI. Machine Learning, Natural Language Processing, and Deep Learning.
Machine Learning
Just like AI, Machine learning is what it sounds. ML is a subset of Artificial Intelligence in which the machine/program is used to learn and train on its own. (Even without human assistance!)
There are a couple of steps that are required for a machine to actually learn.:
Datasets: Machine learning requires tons of data to train on. It uses data to recognize patterns, and using these patterns, it becomes exceptional at performing human tasks over time. At this time, you must have heard that self-driving cars are becoming more of use. Machine learning algorithms make it possible. Using cameras, a lot of data is created about the city/road they are driving through.
Training: Using the data, a ML model runs a program that analyzes patterns and decisions to make a self driving car functional. Every Machine Learning algorithm requires training to make the model as accurate as possible with the data it’s given.
For Machine Learning, there are just a couple things to remember; loads of data and training.
You could think of it as a professional athlete. They would need a lot of information from the coaches on how to get better (data) and then they’d practice for a very long time (training).
Deep Learning
There are countless times where I had to use Google Maps to figure out where a destination is and how I can get to it. The technology behind this is Deep Learning.
Machine learning comes with subsets too.
While machine learning is about programs being able to learn with little to no human intervention, deep learning is quite similar, except it trains on structures based off of the human brain.
These structures are known as Neural Networks.
You can see from the image above that neural network provide multiple layers of classification. In the input layer, we would provide the data we want to train with. For instance, this could be an image that we want to classify.
The output layer on the other hand, gives us the solution to what we classified. This could be telling the difference between a carrot and an apple.
In between, you can see that there are multiple different hidden layers that perform the computations required. The images would be broken down into smaller pixels that would then be transferred to the first hidden layer.
From here, numbers are assigned to each of the layers and with the numerical calculations, neurons are passed on to the following layers, till we get an accurate prediction about the classification.
We can see this being put into play when we perform Facial Recognition.
Neural Networks can estimate the age of a person based off of their features just by corelating lines on the face.
Natural Language Processing
Machine Learning isn’t the only branch of AI that is widely used right now. We see Natural Language Processing (NLP) being used almost every single day. For example: spell check, voice assistants like Siri and Alexa, and many other text and speech related AI programs.
Computers break down text to analyze patterns and the structure of the given piece. From there, the computers are able to read and understand spoken/typed text that is given to them.
This whole process comes down to 4 main steps:
Tokenization: The computer breaks down written or spoken text to analyze patterns and structures. (Usually simple clauses)
Speech Tagging: Think of this as learning English all over again. In this step, the computer detects and breaks down nouns, adjectives, verbs, etc. It marks this so that it can use it for the next steps.
Stemming: Allows computers to standardize the words by converting them to their original forms.
Stop Word Removal: This step is quite similar to the speech tagging step, but instead of marking out verbs and adjectives for example, it marks out common words that add no unique characteristics or importance to the text.
Only after doing these steps can a NLP algorithm transform text into something a computer can easily understand and interpret.
Applications of Artificial Intelligence
There are many different applications of AI as we see today, but there are 3 of them that stood out the most:
Self Driving Vehicles: I truly believe that self driving cars will be the future of commercial road transportation because of its accuracy with AI. AI with the help of Machine Learning can develop self driving vehicles, and we see this already being used around the world with many car manufacturers.
AI In Marketing: I’m sure you have wondered how you get advertisements tailored to what your interests are. Using AI, marketers can target audiences to boost their sales and to ensure better results.
AI In Healthcare: I can guarantee that Artificial Intelligence is revolutionizing healthcare right now. AI is helping in many ways ranging from helping patients navigate the hospital, to even diagnostics by analyzing patterns in blood/tissue samples.
In the future, AI will be in every bit of our lives and we can already see this happening today. AI is not only being used to help peoples needs in a more efficient way, but it is impacting our lives in such a great way.
This is only going to grow even further and change the world.
