Recent Tutorials and Articles
    Introduction to Deep Learning and Neural Networds
    Published on: 25th May 2018
    Posted By: Amit Kumar

    This article will provide basic knowledge of Deep learning and Neural networks along with their use cases

    What is Deep Learning?

    Deep learning, also known as hierarchical learning, is a class of Machine learning and was discovered in 2006. It is broadly defined as set of techniques to train and learn in neural networks that we will talk about in next section. Similar to conventional machine learning, Deep learning can be supervised(models are pre-trained using training data before live deployment) or unsupervised(models are trained as live data comes through). 

    Moreover, Deep learning works by breaking the problem in more than three layers(including input and output layers) of representations(abstractions) wherein each layer uses the output from previous layer.


    What are Neural Networks?

    Neural networks are a set of algorithms, modeled loosely after the human brain, that are designed to recognize patterns. They interpret sensory data through a kind of machine perception, labeling or clustering raw input. The patterns they recognize are numerical, contained in vectors, into which all real-world data, be it images, sound, text or time series, must be translated.

    Neural Networks are also referred as Artificial Neural Networks (ANNs). Here is the general idea of an ANN - 

    Artificial Neural Networks

    As we can see, output of previous nodes is passed as input to successive nodes. Moreover, nodes in each layer of ANN represents a Function (f) that converts input to output. This function could be a logical function (AND, OR etc.), linear function (3*I + b where I is input vector and b is bias) or any other function. This function is derived by ANNs with help of training data and then these trained models are used to get output of set of input features.

    ANNs can be of various types depending on number of layers or direction of data flow as described in below nice diagram by Fjodor van Veen from The Asimov Institute -

    Neural Networks Chart


    Use Cases

    Some of the use cases of Deep learning are as follows -

    1. Computer vision
    2. Speech recognition
    3. Natural language processing
    4. Audio recognition
    5. Social network filtering
    6. Machine translation
    7. Self driving cars


    Libraries and Frameworks

    1. TensorFlow
    2. DeepLearning4j
    3. Theano



    "Neural Network Zoo" by Fjodor van Veen & The Asimov Institute

    Thank you for reading through the tutorial. In case of any feedback/questions/concerns, you can communicate same to us through your comments and we shall get back to you as soon as possible.

    Posted By: Amit Kumar
    Published on: 25th May 2018

    Comment Form is loading comments...