When I started out on machine learning, I had a lot of confusion. My confusion wasn’t anything technical but how words were thrown around on my journey to machine learning.
I heard words like data science, artificial intelligence, machine learning and deep learning. Within these scopes, there are still many words that arouse curiosity.
You might have wondered how they differ. Hopefully, I would be able to clear your doubt, so grab a seat!
These areas are evolving fast and the definition you find here today might be different from what you will find tomorrow, so don’t forget to keep at pace with the growth of technology.
What is artificial intelligence?
Before looking out for the meaning of artificial intelligence specifically, I had a notion that artificial intelligence (AI), was about robots taking over the world by being able to do the same things that we, as humans could do.
While this is part of the truth, this is not entirely what artificial intelligence is all about. As we know, half truth is almost no truth.
The word intelligence according to Merriam-webster dictionary is
‘the ability to learn or understand or to deal with new or trying situations’. It is also defined as the skilled use of reason and the ability to apply knowledge to manipulate one’s environment or to think abstractly as measured by objective criteria (such as tests)
Artificial intelligence (AI) is therefore, based on the idea of the capability of a machine or computer program to think(reason), understand and learn like humans.
From the definition of intelligence, we can also say that artificial Intelligence is the study of the possibility of creating machines able to apply knowledge received from data in manipulating the environment.
Still buzz words? Wait! In simple terms…
AI (artificial intelligence) is reproducing human intelligence in machines, especially computer systems through learning, reasoning and self-correction.
Real life example of AI:
If you are my friend, and I understand that you love action movies , I would make suggestions of action movies to you, based on what I know about you. This is human intelligence.
Machines have also become able to do this, if you watch a particular category of movies on Netflix for example, Netflix starts making movie suggestions to you, based on your watch pattern.
How is this possible? Artificial intelligence. This is a very general example of Artificial intelligence.
What is Machine learning?
Artificial intelligence is very vast. Machine learning(ML) is a subset of Artificial Intelligence. Remember the learning aspect of the definition of intelligence from the definition in the previous paragraph? That is where ML comes in.
Machine learning(ML) is a set of statistical tools to learn from data. The nucleus of ML is in teaching computers how to learn and make predictions from data without necessarily being programmed.
Real life example of ML:
We all get spam mails. These are always filtered out by gmail for example. Also, mails are categorised as promotions and social, as well as other categories based on the mail service you use. How has gmail learnt to do this? Machine learning! Don’t forget ML is part of AI. Over time, the service has been able to learn the category that an email might fall into. It may be wrong sometimes, but it keeps learning.
What is Deep Learning?
In machine learning, data mostly passes through algorithms which perform linear transformations on them to produce output.
Deep learning is a subset of machine learning in which data goes through multiple number of non-linear transformations to obtain an output.
‘Deep’ refers to many steps in this case. The output of one step is the input for another step, and this is done continuously to get a final output. All these steps are not linear. An example of a non-linear transformation is a matrix transformation.
Deep learning is sometimes called deep neural networks(DNN) because it makes use of multi-layered artificial neural networks to implement deep learning.
Seen a photo of a neuron from the human brain? Artificial neural networks are built similarly, with neural nodes connected like a web.
Deep learning algorithms require very powerful machines and is very useful in detecting patterns from input data.
An application of Deep Learning:
Ever heard of WaveNet and Deep Speech? They both are Deep Learning networks that generate voice automatically. Text to voice systems, before WaveNet and Deep speech were manually trained.
With deep learning, systems are learning to mimic human voices to the point where it is hard to distinguish between a human and a computer voice-over. Deep Learning draws us closer to giving computers the ability to speak like humans.
Deep learning is a subset of ML which is a subset of AI, so it is AI.
What is data science?
Data science has an intersection with artificial intelligence but is not a subset of artificial intelligence.
Data science is the study of an aroused curiosity in any given field, the extraction of data from a large source of data related to the question in mind, processing data, analysing and visualising this data, so as to make meaning out of it for IT and business strategies.
In simple terms, it is understanding and making sense of data. A lot of tools are used in data science. They include statistical tools, probabilistic tools, linear and metric algebra, numerical optimisation and programming.
An application of data science :
Pick a random concept.
I choose sponsorship. How do people get sponsorship for a cause. Who is usually willing to respond to an email calling for sponsors. What keywords do they look out for in emails requesting for sponsorship? would they prefer a phone call?
In this case, data science can help. A pool of data related to everyone who has ever sponsored a cause, why they sponsored it, their preferences in terms of communication channels etc is pulled up a large set of unstructured data .
The data is processed, analysed and visualised using the various tools. Conclusions are made from this data.
This information can help non-profits and people pursing a cause to look out for sponsors.
Data Science is not fully Artificial intelligence, however portions of Data science intersect with Artificial intelligence.
When it comes right down to it, one thing is common to these buzz words — DATA !
Top comments (4)
Artificial Intelligence (AI) is a broad field aiming to create systems that can mimic human intelligence. Machine Learning (ML) is a subset of AI focused on teaching systems to learn from data and make predictions. Deep Learning (DL) is a specialized ML technique using neural networks to handle complex data like images and speech. Data Science involves analyzing data to extract insights and often employs AI, ML, and DL techniques. In AI development solutions, these technologies are integrated to build intelligent systems and derive valuable insights from data.
great article
Woah , this is really enlightening!!
Very Informative!