Lorem ipsum dolor sit amet, consectetur adicing elit ut ullamcorper. leo, eget euismod orci. Cum sociis natoque penati bus et magnis dis.Proin gravida nibh vel velit auctor aliquet. Leo, eget euismod orci. Cum sociis natoque penati bus et magnis dis.Proin gravida nibh vel velit auctor aliquet.

  /  Project   /  Blog: Deep Learning Series Chapter 1: Introduction to Deep Learning

Blog: Deep Learning Series Chapter 1: Introduction to Deep Learning

Currently, AI is progressing rapidly and deep learning is one of the contributors. Deep learning is a branch of machine learning that constantly changes the world around us.

From driver-less cars to voice recognition, Deep Learning makes everything possible. It has become a hot topic for industry and science and affects almost all industries related to Machine Learning and artificial intelligence (AI). This is the principal article in deep learning arrangement and will clarify diverse deep learning models in coming articles in the arrangement.

What is Deep learning?

AI vs ML vs Deep Learning

Deep learning is a sub-field of machine learning dealing with algorithms enlivened by the structure and function of the brain called artificial neural networks. As it were, It copy the working of our minds. Deep learning algorithms are like how sensory system organized where every neuron associated one another and passing information. Deep learning is part of a broader family of Machine learning methods based on learning data representations, as opposed to task-specific algorithms. Learning can be supervised, semi-supervised or unsupervised.

Deep Neural Networks

Deep learning models work in layers and a typical model at least have three layers. Each layer accepts the information from previous and pass it on to the next one.

Why should one know about Deep learning?

Deep Learning vs Older Learning Algorithms

One of the major reason for deep learning is the amount of data we handle now a days. Deep learning models tend to perform well with amount of data whereas old machine learning models stops improving after a saturation point.

Nearly every industry is going to be affected by AI and ML and Deep learning play a big role in it. Regardless of whether you are a health professional or a lawyer, there is a possibility that one day you may be replaced by a highly autonomous robot. The accuracy of deep learning has improved significantly over the years and continues to evolve. Understanding its nuances will help us all.

Some of the wide applications of Deep learning are:

Self Driving cars : A self-driving car is the ultimate evolutionary goal of developing ADASes — Advanced Driver Assistance Systems, to the point when there’s nobody to assist anymore.

Visual tasks, which include, among others, the recognition of lanes, the recognition of pedestrians and the recognition of traffic signals, are solved through in-depth learning.
The importance of deep learning for autonomous driving systems can be illustrated by the fact that Nvidia maintains long-term relationships with car manufacturers and works on integrated and real-time operating systems developed for that purpose.

Humanoid: In the same way, deep learning simplifies the interaction between robots and human beings day by day. We already have personal agents like Alexa and Siri who listen to our questions and respond intelligently.

The great advances in NLP and image processing that deep learning has made possible are the reason for this efficient interaction. In view of the growth rate of robotics and deep learning, autonomous robots are not far away. A good example being Google Duplex, a human-like virtual assistant by Google.

Medical care: the adoption of deep learning in health care is increasing and solving a variety of problems for patients, hospitals and the entire health care industry.

Research has shown that deep neural networks can be trained to produce radiological results with high reliability by training archival data from millions of patient scans collected by health systems. These advances will soon change the health and care scenario by replacing doctors with AI-based expert systems and autonomous robotic surgeons.

This should be enough to give you an idea of the vast applications of Deep learning. Unless you’re planning to head in the woods, sooner or later you’ll get to interact with DL in some manner. Now let’s have a look at how it works!

Implementation of Deep learning

Given that Deep learning is implemented by large Artificial Neural Networks (or simply Neural Networks or NN), let’s find out more about them.

What’s an Artificial Neural Network

Artificial Neural Network is a network of interconnected artificial neurons (or nodes) where each neuron represents an information processing unit.

These interconnected nodes pass information to each other mimicking the human brain.

The nodes interact with each other and share information. Each node takes input and performs some operation on it before passing it forward.

The operation is performed by what is called an Activation function (non-linearity). It converts the input into output which can be then used as input for other nodes.


The links between nodes are mostly weighted. These weights are adjusted based on the performance of the network. If the performance (or accuracy) is high, then weights are not adjusted, but if the performance is low, then weights are adjusted through specific calculation.

The leftmost layer of neurons is called the input layer and similarly, the rightmost layer is called the output layer. All the other layers in between are called hidden layers. In a nutshell, an Artificial neuron takes input from other nodes and applies the activation function to the weighted sum of input (Transfer function) and then passes the output. A threshold (called Bias) is added to the weighted sum to avoid passing no (zero) output.

A Neuron

For knowing more about Neural Networks, check NEURAL NETWORKS by Christos Stergiou and Dimitrios Siganos.

How are Neural Networks used for Deep learning

For Deep learning, several Neural Network layers are connected in feedforward or feedback style to pass information to each other.

Feedforward: This is the simplest type of ANN. Here, the connections do not form a cycle and hence has no loops. The input is directly fed to output (in a single direction ) through a series of weights. They are extensively used in pattern recognition. This type of organization is referred as bottom-up or top-down.

Feed Forward

Feedback (or recurrent): The connections in feedback network can move in both directions. The output derived from the network is fed back into the network to improve performance (loops).

These networks can become very complicated but are comparatively more powerful than feedforward. Feedback networks are dynamic and are extensively used for a lot of problems.

Now let’s discuss some specific types of ANN extensively used for Deep Learning. All these types will be discussed in detail separately in upcoming articles.

1). Multilayer Perceptrons: These are the most basic Neural Networks with feed-forward networks. They generally use non-linear activation functions (like Tang or Relu) and compute the losses through Mean Square Error (MSE) or Logloss. The loss is back propagated to adjust the weights and make the model more accurate. They are generally used as a part of a bigger deep learning network.

Artificial Neural Network

2). Convoluted Neural Network: Convoluted Neural Networks (ConvNet or CNN) are similar to ordinary Neural Networks but their architecture is specifically designed for images as input. In particular, unlike a regular Neural Network, the layers of a ConvNet have neurons arranged in 3 dimensions: width, height, depth.


They are particularly suitable for spatial data, object recognition and image analysis using multidimensional neurons structures. One of the main reason for the popularity of the deep learning lately is due to Convoluted Neural Networks. Some of the common usages of Convoluted Neural Networks are self-driving cars, drones, computer vision and text analytics.

3). Recurrent Neural Networks: RNNs are also a feedforward network, however with recurrent memory loops which take the input from the previous and/or same layers (backpropagation). Here connections form a directed graph along a graph. This gives them a unique capability to model along the time dimension and arbitrary sequence of events and inputs.

Image result for rnn intuitive picture

In simpler terms, for any given instant, the network maintains a memory up till that moment and therefore can predict the next action. Most common types of RNN model is Long Short Term Memory (LSTM) network. RNNs are used for next work prediction and grammar learning.

This post aimed at providing a brief introduction to the massive field of Deep learning. I have skipped mathematical details of all the concepts, which will we covered in the separated and detailed stories of all the above mentioned concepts.

Source: Artificial Intelligence on Medium

(Visited 11 times, 1 visits today)
Post a Comment