TRENDING ON ONEINDIA
- Ananth Kumar Passes Away: PM Narendra Modi Flies Down To Bangalore
- Here Are all The Upcoming Smartphone Launches For The Month Of November 2018
- Maruti Electric Cars From E-Trio Automobiles: Electric Alto And WagonR
- Nick Jonas Bachelor's Party: New Pictures Emerge
- How To Turn Any Financial Crisis Into A Profit Making Opportunity?
- Migrane? Here Is a List Of Diseases Linked To Migraine
- Must-know Interesting Facts About Kochi
- ICC rankings: Kuldeep, Dhawan & Hafeez Move Up The Charts
Artificial Intelligence, as we know is the future, fiction, and a part of our daily lives. Yes, that's right, it just depends on what type of AI you are referring to. You might come across the term 'AI' quite frequently. We see it in movies, Google Search, chatbots in messengers, and not to mention the theories where the AI will take over the world.
While the Artificial Intelligence has been around for quite some time now, there have been new technologies paving the way into the mainstream. Machine Learning and Deep Learning are bursting on to the scene, with many still wondering what exactly both of them are. Artificial Intelligence, Machine Learning, and Deep Learning together can do things beyond imagination. But there's a difference among the three.
The simplest way to understand the connection between these three is to visualize them as concentric circles with Artificial Intelligence being the largest, then Machine Learning which came later, and lastly Deep Learning which is responsible for the sudden boom in AI industry. Let's learn more about them.
What is Artificial Intelligence
Artificial Intelligence is the general most category which is used in the application of all the three technologies. There are many ways to simulate human intelligence. While some are very basic, others are more intelligent than others.
AI can be a load of if-then statements or a complex statistical model that maps raw sensory data to symbolic categories. The statements would be simple rules created by a human. These if-then statements are also referred rules engine, expert systems, knowledge graph of symbolic AI. All of these aspects come together to form an AI.
Artificial Intelligence is basically any sign of intelligence shown by a machine that results in an optimal or suboptimal solution to an issue. The simplest form of AI can be seen on Tic-Tac-Toe AI player. If an AI bot works on a specific pre-programmed algorithm, it will never lose a game.
So the basic concept of Ai is to develop something based on advanced algorithms allowing it to make its own calls, or simply do something which computers can't.
What is Machine Learning
Machine Learning is the most common practice that uses algorithms to parse data, learn from it, and based on the data, make a prediction about something in the world. So instead of hand-coding software routines, and following a specific set of instructions to perform a task, the machine is taught how to use huge amounts of data and algorithms giving it an ability to learn how to execute a task.
For instance, a house price prediction model goes through a lot of data, with each data point comprising numerous dimensions such as size, bathroom count, yard space, and others. It creates a function using these inputs, and then just shifts the coefficients to each of these parameters as it looks at more and more data.
This process is called "Supervised Learning," where the data provided to the model has the answers to the problems for each input set. It's actually providing the input parameters, and the outputs for every single set of features, using which the model adjusts its function to match data. Apart from this, there exists Unsupervised Learning and Reinforcement Learning.
Unsupervised Learning just finds similarities in data.
Machine Learning is overall an optimization algorithm. If you tune it right, they minimize the margin for error by predicting and predicting and predicting again.
What is Deep Learning
This is a state of the art technology. As for now at least. Deep Learning is actually a subset of Machine Learning. Usually when people use the term "Deep Learning," they are referring to deep artificial neural networks and deep reinforcement learning.
These artificial neural networks are a set of algorithms that have par excellent in terms of their accuracy for many problems like image recognition, sound recognition, recommender systems, etc. The best example would be DeepMind's well-known AlphaGo algorithm, which beat the former world champion Lee Sedol at Go in early 2016, and the current world champion Ke Jie in early 2017.
It refers to the number of layers in a neural network. A shallow network has one so-called hidden layer, and a deep network has more than one. One of the other hallmarks of deep learning is the computational intensivity. It is also one of the prime reasons why GPUs are in demand to train deep-learning models.
So Deep Learning can be defined how Arthur Samuel defined Machine Learning - a "field of study that gives computers the ability to learn without being explicitly programmed," but with higher accuracy and low training time. It also performs exceptionally well on machine perception tasks that have unstructured data such as texts or Pixel blobs.