1. Algorithms and AI systems demystified
Hi, I'm Maarten, the second guide for this course.
This chapter will further explore the tasks AI performs and the areas and techniques involved.
But first, let me help you understand and distinguish between two key concepts that are often used interchangeably in AI-related media: algorithms and AI systems.
2. What is an algorithm?
An algorithm is a key element not only of AI but of computer science as a whole. It can be defined as a set of computer instructions, rules, or operations to solve a problem or perform an action.
An algorithm has three main parts. The inputs are the data or information needed to initiate the process.
3. What is an algorithm?
The process is the set of operations or instructions to perform. For instance, if our algorithm calculates the average height of a large group of people whose measurements are passed as inputs that mix heights expressed in cm and inches, the algorithm's steps would include normalization into a common measurement followed by calculating the average height.
4. What is an algorithm?
And the output is the result of applying the process or the solution for the problem.
You can think of an algorithm as something that resembles preparing a recipe like a banana bread. The inputs would be the ingredients.
5. What is an algorithm?
The process includes the steps for preparing the dough and baking.
6. What is an algorithm?
And the resulting output is a delicious loaf of banana bread.
7. Algorithms in Computer Science vs AI algorithms
The boundary between what is a conventional computer science algorithm and what is an AI algorithm might not seem too clear at first glance.
8. Algorithms in Computer Science vs AI algorithms
For example, a computer science algorithm is what happens in a pizza store every time an order is placed, calculating the price of a customers order depending on the food and drink chosen, the price list, and any promotional codes used.
9. Algorithms in Computer Science vs AI algorithms
Meanwhile, the process to generate personalized recommendations for a customer in a hotel booking portal is a typical example of an AI algorithm. This process involves more complex steps than the rules and operations to calculate the price of an order, such as analyzing and modeling customer preferences and estimating customer-to-product affinity.
As a rule of thumb, AI algorithms have the ability to learn to produce better outputs or to improve their processes by themselves, adapting to the input data; whereas conventional computing algorithms have a more rigid and deterministic input-output behavior.
10. What is an AI system?
Another question to raise is: are AI algorithms and AI systems the same?
The short answer is no. We already know an algorithm is a procedure or set of instructions to perform a task or reach a goal.
11. What is an AI system?
By contrast, an AI system encompasses the broader infrastructure and components needed to implement and deploy our AI algorithm into the real world.
12. What is an AI system?
This includes elements like hardware and software, database and data storage, network infrastructure, and so on.
Back to the AI algorithm example for personalized hotel recommendations, the equivalent AI system would include elements like the customers, hotels, bookings and ratings database, the underlying data, network, hardware infrastructure, etc.
13. Let's practice!
Now we will check what we learned about algorithms and AI systems through some practical exercises.