Uses of AI

Logic

We already discussed the uses of AI on the introduction page to some extent. On this page, we will take a deeper look into the applications of AI and how AI functions in each of these use cases. However, we must learn one last concept first: logic programming.

The best weapon of any AI is logic, and an AI can gain this ability through logic programming. Logic programming is a programming paradigm that uses simple rules of logic to solve problems. To solve a problem using logic programming, you need to identify these three things: the facts, the rules, and the goal.

• The facts are simply any correct statements found within the problem.
• The rules are the restrictions that allow the program to conclude (e.g., if A=B and B=C, then A=C).
• The goal is the conclusion we will arrive at using facts and rules.

This programming paradigm is best suited for solving logic problems, which are problems AI frequently encounters (as you shall see later). With that out of the way, we can now proceed to the applications of AI.

Data Science

Data science is a field that utilizes statistical techniques along with AI to extract value (e.g., predictions and generalizations) from data. Its purpose is to analyze large datasets that no one can analyze by themselves. Say, all the tweets ever sent, or all the Facebook posts ever made. A data science project flows much like an AI project. You start by determining the project's purpose, then programming the machine learning algorithm, then evaluating the algorithm, and finally, deploying the project.

Because of their nature, data science and machine learning go hand-in-hand . Machine learning aids in data science through sequence analysis. Sequence analysis allows an AI to predict missing data and estimate future values, thus generating a more comprehensive dataset. Meanwhile, data science may serve as a tool for preparing an AI's training dataset. Thanks to the power of data science software to process large datasets, AIs can train with more extensive and higher-quality datasets.

NLP

Natural language processing is one of the uses of AI most people encounter every day. For example, you have predictive text, google translate and text-to-speech software. Essentially, natural language processing is a method of making computers understand human languages through AI. It has two main components: natural language understanding and natural language generation.

Natural language understanding is a process wherein an AI attempts to understand text or speech input by breaking it down into words and determining each word's purpose. For example, let us simulate what will happen if we use natural language understanding on the sentence "I love you!" The results will state that "I" is the subject, "love" is the verb, and "you" is the object.

Once the input has been broken down and understood by the AI, the AI may need to generate an appropriate response, which is what natural language generation does. Natural language generation works the other way around compared to natural language understanding. It starts by picking words that form a meaningful sentence, then determining the purpose of each term, then arranging the words into a string according to their purpose. If we give the AI an input equivalent to the example above, then most likely, the output we will get would be "I love you too!"

Gaming

AI also has its uses in the gaming world. Say, non-player characters, or NPCs, which make games more exciting. Under the hood, AI drives these NPCs. These NPCs function through algorithms that find the shortest path towards a goal, otherwise known as search algorithms. There are many different ways to implement search algorithms, but these methods usually involve something known as heuristics. Heuristics is a method of solving problems where solutions that are less likely to result in a correct answer are discarded and not processed. Heuristics can save computer resources when we run a search algorithm.

Personally, I (the author) would recommend you start your AI journey here, as it is the easiest to implement. For example, you can create a tic-tac-toe bot using Python in less than 30 lines of code. All you have to do is use a search algorithm known as the Negamax algorithm, then define the size of the board, the rules of the game, and the winning conditions. From there, it is just a matter of taking input and printing output, and you will have your bot!

Computer Vision

Finally, we have one of the most recent applications of AI: Computer Vision. Computer vision allows computers to understand and interpret visual media, such as images and videos. There are three levels to computer vision:

• Low-level - This allows a computer to detect the general features of an image, such as colors, edges, and lighting.
• Intermediate - This allows a computer to spot specific structures within images, such as faces, eyes, and cars.
• High-level - This allows a computer to predict the movement of objects in its view.

Though computer vision is still a relatively new technology, it already has a lot of applications we can see in our daily lives. For example, some apps use computer vision to detect your face and apply a filter. Self-driving cars also use computer vision to determine which lane to drive on. The police can also use computer vision to catch a criminal on the run, and the list goes on.