This is a post from series on artificial intelligence and machine.
In this post, we will try to understand what AI is and where machine learning fits in it.
As per Wikipedia Artificial Inteligence is
Simulating any intellectual task.
It can be also seen as the industrial revolution to simulate the brain.
AI is a very broad field and it contains many subfields and it is important to understand what the full landscape looks like and focus on the core part that overlaps with almost every subfield of AI.
Let's try to understand each subfield.
Knowledge representation
This is core to many AI applications, it is based on an expert system that collects explicit knowledge that is available in some database or possessed by experts.
This can be also seen as Knowledge about knowledge. We interact with system type of system every day be it Amazon Alexa, Apple Siri, or Google Assistance.
Perception
Machine Perception is about using sensor input to understand context and action to take. Nowadays we are surrounded by cameras, microphones, IoT devices, etc.
Some real-world applications include facial recognition, computer vision, speech, etc.
Motion and manipulation
This is one of the heavy use of AI, it includes robotics. The industrial revolution has already helped the world economy grow, and robotics will take it to the next level. Some applications in industrial/domestic robots. In the time of pandemics like Covid, robotics is even going to help more as everyone is concerned about safety. Autonomous vehicles are one of the important applications of this sub-field.
Natural language processing
NLP allows the machine to read and understand human language. It includes processing huge unstructured data and derives meaning from it. Some of the application that we get interact every day is search autocomplete, auto-correction, language translator, chatbots, targeted advertisement, etc.
Search and planning
This area covers machine that is set a goal and achieves it. The machine builds the state of the world and can make predication on how their action will change it.
Learning
This is also called as Machine Learning and it is the study of computer algorithms that automatically improve through experience.
It sounds like how humans learn something!
It is a subfield of AI but the most important one as it is applied to all the subfields of AI, knowing this is a must before starting on any other subfield of AI.
Let's explore more on the Learning part now.
What is machine learning?
One of the quick definitions of machine learning is pattern recognization, it can also be seen as how computers can discover to solve problems without explicit programming.
Machine learning is made up of 3 steps.
The step of updating the model via learning is where real machine learning happens.
Data science is related to machine learning but is often seen as only machine learning. AI & data science good overlap with machine learning, it can be seen as below.
 |
| Data Science |
Now with a high-level understanding of AI, ML & data science, we are ready to do deep dive in ML.
