Deep Learning series

Chapter : 1

Mahesh Patel(Data Science Researcher)

Jan 30 2021

AIM : Aim of this article is to know about Deep Learning and Why is Need of Deep Learning in AI.

WHAT is Deep Learning :

Deep learning (also known as deep structured learning) is part of a broader family of machine learning methods based on artificial neural networks with representation learning. Learning can be supervised, semi-supervised or unsupervised

Deep-learning architectures such as deep neural networks, deep belief networks, recurrent neural networks and convolutional neural networks have been applied to fields including computer vision, machine vision, speech recognition, natural language processing, audio recognition, social network filtering, machine translation, bioinformatics, drug design, medical image analysis, material inspection and board game programs, where they have produced results comparable to and in some cases surpassing human expert performance.

Deep Learning is performed using Artificial Neural Networks.

WHY we need Deep learning :

Data Size vs Model Performance

1. Classical ML model do not improve after a certain point no matter if you increase data size and no of features.
2. ML models fails to model complexity of data after a certain point.
3. Non Linear complexity is far better handled by Deep Learning.
4. And By above accuracy graph vs data size proves Deep Learning/Neural Network is far better when you have a large no data points.

Advantages of Deep Learning :

1. Best Accuracy if modelled properly.
2. Complex problems are modelled easily compared to Machine Learning.
3. You don’t need to do feature engineering and modelling differently as in machine Learning.
4. So its Easy, Fast, and Accurate to make a Deep Learning model compared to a Machine Learning Model.

Disadvantages of Deep Learning :

1. Lots of Hyper-Parameters to Fine Tune.
2. If you don’t understand basics of Layers, Neurons, Activation functions, Optimization Techniques and all, it will be hard to make or understand a Neural Network.
3. Because of so much of calculation the computation cost is very much high as compared to Machine Learning.

Conclusion :

If you Have large amount and complex data then you should opt for Deep Learning/ Neural Network.

Please do give your Suggestions.

References :

………………………………………