Overview
- A machine learning system consists of several basic components that must be managed
- Learn about the three key building blocks of machine learning that you will work with as a data scientist
Introduction
How does a machine learning project work? What are the different building blocks that go into creating a machine learning or artificial intelligence system? (HE)? This is an issue that I personally struggled with during my early days on the pitch..
He knew how to make machine learning models, but i had no idea how a real world machine learning project actually worked. It was a great revelation when I went through the process!! And with the time, I've seen most data science and machine learning beginners struggle to understand the nuances of a machine learning system.
Remember: It's not just about building models! There is a LOT it takes to create a successful machine learning and artificial intelligence system. It is a fusion of hardware and software, among other things. Then, the question is: What are the key building blocks that make up a successful machine learning system?
That is what we will cover in detail in this article.. I will give you an overview of these different components in a machine learning or artificial intelligence system., and then we will understand these components with the help of an autonomous car.
This article and the concepts we will cover are part of the free program ‘Introduction to AI and Machine Learning‘ course. I highly recommend checking it out: it's a great place to get acquainted with the various concepts of artificial intelligence and machine learning.
And the three fundamental pillars of machine learning are:
- Machine learning building block n. ° 1: capture input
- Machine learning building block n. ° 2: data processing and storage
- Machine learning building block n. ° 3: unit of outcome or interaction
Machine learning building block n. ° 1: capture input
As expected, every machine learning system needs a large amount of data to function. As a last resort, make decisions based on the data you capture. And you need to capture data about the environment you are in, environmental conditions, user inputs, etc.
Therefore, the first building block of any artificial intelligence or machine learning system is the way it captures and enters the system.
Then, How does this entry look? This could include various sensors such as a camera that captures images, GPS location, user inputs from mobile apps, etc. To select the correct entries, we must ask these key questions:
- What data do we need to capture?
- How often do we need to capture this data?
- How fast would this data flow?
- What could be the best way to capture this data?
Sometimes, there would be several ways to capture the same data. For instance, you can rely on your car's sensors to capture weather information, or you can pull them directly from the internet based on your car's GPS coordinates.
It might make sense to weigh the pros and cons of various ways of capturing data before deciding which one you prefer..
Machine learning building block n. ° 2: data processing and storage (Edge y Cloud)
Once we capture this data from the input units, we will need to store them or run calculations on them. That is basically the choice it comes down to when we are working on a machine learning project!
Both of them (processing or storage) can occur in the system normally called “AI on the edge” or they can happen in the cloud. Again, we have some options in front of us. We need to decide:
- What data would be stored on the edge?
- What calculations would occur on the edge? Here, would normally have limitations in the computing environment (cream, Not everyone has Google's unlimited computing resources!!)
- What would happen in the cloud?
As usual, if there is any critical operation that should occur, even if there is no internet connection or a system update, should always occur on the edge.
These would include things like decisions on the fly, alerts or any other form of monitoring you want on the device. The most complete data storage and calculations are done in the cloud. This is where data scientists often apply various machine learning techniques to improve the system.. All our data lakes, data warehouses, etc. normally they would also be in the cloud.
Machine learning building block n. ° 3: unit of outcome or interaction
Finally, there would be a unit of output or interaction in a successful artificial intelligence or machine learning system. This is the unit in which the machine learning system would interact with the outside universe and act.
This could be in screen form, voice output or informal robotic actions. As usual, the output of our machine learning system would also have various design considerations.
For instance, if a vehicle cannot decide or read the environment with certainty, key questions need to be answered:
- What should the system do?
- Should it stop first or should it alert the user?
- How often and what details should you communicate to the user?
Estas son algunas de las preguntas centrales que se consideran comúnmente en la Output layerThe "Output layer" is a concept used in the field of information technology and systems design. It refers to the last layer of a software model or architecture that is responsible for presenting the results to the end user. This layer is crucial for the user experience, since it allows direct interaction with the system and the visualization of processed data.... de cualquier sistema de aprendizaje automático.
Case study: building blocks for an autonomous car
Now, Let's take an example of a self-driving car and look at each of these building blocks in more detail.. This will help you gain a more practical understanding of how a machine learning or artificial intelligence system works in the real world..
Then, What would be the first building block or component?
You guessed it, entry! Check out the autonomous car from drive.ai:
As you can see here, this autonomous vehicle has many sensors that act as input to the machine learning system. You can see these sensors on the top of the car (In color blue). These are called LiDAR or light sensing and range. Besides these, there are other sensors that capture more information such as the weather, obstacles in and around the car, lane detection, etc.
Then there is the calculation and storage in the car itself, allowing you to make decisions like:
- How much to drive?
- How fast to run?
- What are the obstacles in the way?
- How to handle these obstacles?
There is also a cloud computing and storage layer, which is responsible for improving the driving algorithm over time.
And finally, I would see various output components as a screen to show messages to people around the car. There is also the action taken by the robotic process to propel the vehicle forward.. Here is an illustration of the different layers that are required at this stage:
There are many other practicalities required to create a successful autonomous car. But I wanted to take this example to show you how the general idea behind a real world machine learning system works and the key building blocks required to run it..
Final notes
Quite fascinating, truth? As I said before, a machine learning project is not just about building models. There's so much more that most data science enthusiasts don't know. This practical knowledge is necessary if you want to obtain a position as a specialist in machine learning.
Here you have a challenge. Now that you understand the various components of an autonomous car, it's your turn to design the components of a smart vacuum that can navigate the floor on its own and clean the area it navigates. Have fun building it!
And if you have questions or thoughts on this article, I would love to hear from you.
