This article was published as part of the Data Science Blogathon.
The range of applications has changed the facets of technology in all fields, from medical care, the making, business, The education, to the bank, information technology and other things.
Although these words are familiar and widely used, are often used interchangeably. But there is a big difference between all of them.
In this article, we will explore these buzzwords and learn the difference between them.
Artificial intelligence
In a nutshell, “artificial intelligence is the ability of machines to function like the human brain”.
Whenever we think of Artificial Intelligence, the first thing that comes to mind are Robots. A few decades ago, ‘Robots’ fascinated us more with movies showing Robots / Superhumans performing incredibly difficult jobs effortlessly and living on a par with humans. Now robots like Sophia are a reality and we find AI everywhere. From robotic vacuums, virtual assistants like SIRI, robots that perform surgery in healthcare, robots that write codes and, of course, autonomous cars and trucks, most of these are a reality and the world of artificial intelligence is rapidly evolving. Starting with the IBM 'Deep Blue' chess computer, who won a game of chess against the World Champion, to 'AlphaGo by Google', we've seen fascinating discoveries in this artificial intelligence revolution.
In simple terms, Artificial Intelligence is about training machines to mimic human behavior, specifically, the human brain and its thinking skills. Like the human brain, artificial intelligence systems develop the ability to rationalize and perform actions that have the best chance of achieving a specific goal.
Artificial intelligence focuses on realizing 3 cognitive abilities like a human: learning, reasoning, Y autocorrect.
The evolution of Artificial Intelligence is considered the fourth Industrial Revolution ~ UBS
Experts say, like the 3 first industrial revolutions changed the course of the world, the fourth revolution, powered by Artificial Intelligence, IoT and the Cloud will surely change the course of humanity and our planet Earth. And enthusiasts run to learn Online AI Course.
Let's take a quick look at the three general categories of Artificial Intelligence and how we are rapidly evolving in these areas!!
1. Narrow artificial intelligence
2. General artificial intelligence
3. Super artificial intelligence
Narrow artificial intelligence
Narrow artificial intelligence Systems are designed and trained to complete a specific task and, often, are called weak AI / Narrow AI. Chatbots that answer questions based on user input, voice assistants like Siri, Alexa and Cortana, facial recognition systems, artificial intelligence systems searching the internet, are examples of weak artificial intelligence. They are smart in perform specific tasks who are programmed to do it.
Narrow AI does not mimic human intelligence, rather it simply simulates human behavior based on a set of input data and parameters. Weak AI still requires a certain amount of human intervention in terms of defining parameters for learning algorithms., feed relevant training data and ensure prediction accuracy.
You can think of him as a baby who listens to instructions from adults and performs functions as directed.
Artificial general intelligence
Artificial general intelligence is when the systems / AI machines act on a par with another human. This also means the ability of the machine to interpret and understand human tone and emotions and act accordingly.. This is also called Strong AI and we're still scratching the surface of Strong AI. As machine learning capabilities continue to evolve, artificial intelligence will progress and we will get there soon.
Super artificial intelligence
Super artificial intelligence / Super AI is when an artificial intelligent machine would become be self-aware and exceed human intelligence and ability. Although a lot of interesting research is being carried out in this area, there are also warnings from scientists.
Oxford University professor and New York Times bestselling author of the book. “Superintelligence: roads, dangers, strategies”, Dice Nick Bostrom,
"The biggest threat is the long-term problem, that introduces something radical that is super smart and doesn't align with human values and intentions. This is a big technical problem. We were able to resolve the capacity issue before we were able to resolve the safety and alignment issue “.
Then, What do you think about reaching this level of artificial superintelligence? Do you think uncontrolled Super AI can become a threat to humanity? Don't forget to share your thoughts in the comments..
Okey! That's a lot about Artificial Intelligence!!
How we have a fair idea of Artificial Intelligence; Wondering how the system / computer machine can imitate human actions and make predictions, automation and decision making?
That's where machine learning comes in..
Machine learning
Machine learning is, of course, a subset of artificial intelligence. Provides statistical algorithms and methods and allows machines to / computers automatically learn from your previous experiences and data and allows the program to change its behavior accordingly.
Machine learning provides many different techniques and algorithms to make the computer learn. Decision trees, Random forests, Support Vector Machines, K means grouping – these are just to name a few.
Demand forecasting product sales, predict customer behavior, measure customer feelings based on their behavior on social media: these are some use cases where machine learning models are used. Machine learning algorithms work well when the input data is reasonably good. When data size skyrockets, we need to look for more efficient algorithms and techniques and that is where Deep Learning finds its access point.
OTT platforms like Netflix and Amazon Prime use Machine Learning to recommend movies based on the user's previous viewing data and constantly improve by learning from past experiences.
Image source: Scrabbl
In e-commerce, Companies like Amazon and Flipkart use machine learning to understand user preferences and offer product recommendations based on previous purchases and viewing history..
The application of machine learning in the real world is huge!!
Now we have a clear idea that machine learning and artificial intelligence are not the same. Machine learning is one of the ways to achieve artificial intelligence.
Like anything else, machine learning has its shortcomings and that's where deep learning comes in.
Machine learning models do not perform very well when the volume and complexity of the data multiply. They need some kind of human intervention and guidance, while deep learning models learn from data and previous experience and are progressively corrected.
Scale with the amount of data | Image source: soshace.com
Let's look at the next buzzword: Deep learning!
Deep learning
Deep learning can be considered as the evolution of machine learning that is inspired by the functioning of the human brain. Deep learning is used to solve complex problems where the data is huge, diverse and less structured. Deep learning models are built on artificial neural networks, that mimic how the human brain works.
Let's look at the basic functioning of our brain to understand how neural networks work. Our human brain has neurons that are the basic functional units of our brain.. Neurons transmit information to other nerve cells, muscles and glands and also receive information from other neurons, allowing the brain to make decisions.
Imagine, you are a small child and you see a pot full of hot water. When you see it, you won't understand it's hot unless someone says so. If you touch the hot pot, the nerves in your fingers carry this information to the brain, and the neurons process this information and send the signal to your fingers and you would feel the heat. Next time I see a hot pot, your brain will remember the previous incidents and remind you that it will feel hot if you touch it.
And our brain continuously learns from the contributions of the environment and previous experiences and makes the best possible decision in each scenario.. This is pretty much what deep learning does!! Learn progressively from raw data and previous experiences and correct yourself without explicit programming.
Image source: DukeToday
Artificial neural networks, convolutional neural networks and recurrent neural networks are some of the deep learning algorithms used to solve real world problems.
Autonomous cars and trucks, virtual assistants like Alexa, Siri y Google Assistant, voice recognition systems, computer vision, robotic surgeries are all cool deep learning applications.
Although Deep Learning was conceptualized in the decade of 1980, the researchers had two major limitations when it came to implementing deep learning models. Deep learning models require a lot of data Y very high computational power. As data grows, the depth of the neural network increases and learning becomes 'Deep'. That is the essence of Deep Learning. Another significant advantage of Deep Learning is that, as the model trains, it learns to extract features on its own and we don't have to perform a manual feature extraction as is the case with other Machine Learning algorithms.
With the advent of new processing units, increased computational power and exponential data growth; Deep learning is gaining momentum and is being used to solve many real-world problems.
“Early in 2020, the number of bytes in the digital universe was 40 times greater than the number of stars in the observable universe” ~ World Economic Forum
In summary, many artificial intelligence systems are powered by machine learning and deep learning algorithms. Artificial intelligence is achieved through machine learning and deep learning and they are not the same.
Hope this article has given you a clear idea of artificial intelligence, machine learning and deep learning.
If you want to leave your thoughts or connect with me on LinkedIn, here is the link- LinkedIn
Happy learning!
The media shown in this article is not the property of Analytics Vidhya and is used at the author's discretion.
