You may be daunted by this and think it’s something that can be done only by top companies and research institutions. If you think so, you are absolutely wrong. In fact concepts like this could be implemented using deep learning by everyone and its fun!
Are you interested in doing what is pictured above? It seems interesting! Isn’t it?
Predicting the future isn’t magic, it’s artificial intelligence.
The thought may be appalling but it is true.
Deep learning is a platform where machine works like a human brain. Now you may be thinking how human brain can be placed in a computer or in any machine?
Deep Learning is human brain embedded in machines. It is one of the most robust branches of Machine Learning. Also, known as Deep Structured Learning and Hierarchical Learning, it can be thought as the simplest way to automate the predictive analysis.
Machine intelligence is the last invention that humanity will ever need to make.
Let’s understand what is deep learning through an example:
Think of an infant who starts learning the alphabets. Initially, mother teaches his/her to learn the alphabet ‘A’ in a standard writing. Gradually the child learns about various types of styles. As the child continues to point to different styles, he/she becomes more aware of the features that all ‘A’ possess. The infant learns the concept of ‘A’ by building a hierarchy using the previous level’s knowledge.
Let’s take a couple of definitions for Deep Learning.
A sub-field within machine learning that is based on algorithms for learning multiple levels of representation in order to model complex relationships among data. Higher-level features and concepts are thus defined in terms of lower-level ones and such a hierarchy of features is called a deep architecture.
— Deep Learning: Methods and Applications
This is what the concept of Hierarchical learning is. A deep learning model tries to learn step by step by itself without supervision instead of all at once (as done in traditional models).
It looks at the input piece by piece and makes lower level patterns from it. Then it uses these lower levels to gradually identify higher level features using many layers, hierarchically.
This hierarchy is used to gradually develop the complicated patterns using simpler ones. This is what helps in seeing the world in a pretty better way. This hierarchy is looked upon not only for features, but also how those features are built in the hierarchy, i.e. it focuses on how the things are made which we call Self Learning in simple words.
This self-learning is the main advantage of deep learning as the model is learnt by itself and not only on what it is taught but what it has learnt throughout the process.
And if the model has hierarchy that means it must possess multiple layers in it to attain maximum accuracy. It is not that we call deep models as Deep Learning. It is that, in order to achieve hierarchical learning the models need to be deep. The deepness is a by-product of implementing Hierarchical Feature Learning.
Deep Learning has become the state-of-the-art in today’s scenario. It is continuously evolving and that day is not far away when humans will completely be replaced by machines. Deep Learning is something really amazing and fun! If you are reading this, then I will advise you to continue in Deep Learning.
You will read Applications of Deep Learning in our next blog.