Machine Learning and Deep learning have become center of attraction now a days in the developer community and the tech industry. Most of the researchers are willing to work on this area. So, the main question comes in mind is, Whether both these techniques are same? So, Let us try to answer this question here. Machine Learning and Deep learning are very much related to one another. So before knowing the difference between machine learning and deep learning it will be better to understand both the terms.
What is Machine learning?
As per wikipedia, "Machine learning is the scientific study of algorithms and statistical models that computer systems use to perform a specific task without using explicit instructions, relying on patterns and inference instead. It is seen as a subset of artificial intelligence".
What is Deep Learning?
As per wikipedia, Deep learning (also known as deep structured learning or hierarchical learning) is part of a broader family of machine learning methods based on artificial neural networks. Learning can be supervised, semi-supervised or unsupervised.
The analogy to deep learning is that the rocket engine is the deep learning models and the fuel is the huge amounts of data we can feed to these algorithms.
– Andrew Ng
[Figure 1: Relation between AI, ML and DL]
As shown in figure 1 Machine learning is the subset of Artificial Intelligence and Deep learning is subset of Machine Learning. So AI? ML? and DL? Now we will move step by step to understand all three.
[Figure 2. ML v/s DL]
As shown in figure 2, In Machine learning feature extraction is done manually while in Deep Learning feature extraction is done automatically as it is the part of its architecture.
[Figure 3: Image of Dogs and Cats]
Now to understand the exact difference between machine learning and deep learning let's take a look at the image above. What you will see is a collection of pictures of cats and dogs. Now if we wish to identify the images of dogs and cats separately with the help of machine learning algorithms and deep learning networks, how can we do that? If we try to define the problem statement it will be as below:
Problem: Identify dogs and cats separately from Figure 3
Solution1: Using Machine Learning Algorithms
To assist the ML algorithm for categorizing the images in the collection according to the two categories of dogs and cats, you will need to present to it these images collectively. But how does the algorithm know which one is dog and which one is cat?
The answer to this question, as in the above definition of machine learning for beginners, is structured data. You have to simply label the pictures of dogs and cats in such a way that can define specific features of both the animals. This data will be sufficient for the machine learning algorithm to learn, and then it will carry on working based on the labels that it understood, and will be able to classify millions of other pictures of both animals cat and dog as per the features it learned through the said labels.
Solution2: Using Deep Learning
Now, if we try to solve the same problem using deep learning networks then it would take a different approach to solve this problem. The core benefit of deep learning networks is that they do not necessarily need structured or say labelled data of the pictures to classify the two animals. The artificial neural networks using deep learning send the input (the data of images) through different layers of the deep neural network, with each network hierarchically defining specific features of images. This is very similar to how our human brain works to solve the problems: by passing queries through various hierarchies of concepts and related questions to find its answer.
After the data is processed through layers within deep neural networks, the system finds the appropriate identifiers for classifying both animals (cats and dogs) from their images.
Difference Between ML and DL
Below we have listed down the major differences between Machine learning and Deep learning:
|
Features
|
Machine Learning
|
Deep Learning
|
|
Amount of data
|
Works good with small amount of data
|
Needs large amount of data to perform well
|
|
Hardware dependencies
|
can work on low-end machines
|
heavily depend on high-end machines like GPU
|
|
Feature engineering
|
Most of the applied features need to be identified by an expert and then hand-coded as per the domain and data type.
|
try to learn high-level features from data
|
|
Problem Solving Approach
|
It is generally recommended to break the problem down into different parts, solve them individually and combine them to get the result.
|
In contrast advocates to solve the problem end-to-end.
|
|
Execution time
|
Comparatively takes much less time to train, ranging from a few seconds to a few hours.
|
Takes a long time to train. In testing, deep learning algorithm takes much less time to run.
|
|
Interpretability
|
Machine learning algorithms like decision trees give us brittle rules as to why it chose what it chose, so it is particularly easy to interpret the hard core reasoning behind it.
Thus, algorithms like decision trees and linear or logistic regression are primarily used in industry for interpretability.
|
Suppose we decide to use deep learning to give automated rating to essays. The performance it gives in rating is fairly excellent and it is near to human performance.
But there is a dispute. It does not reveal why it has given that score or say rating.
Certainly mathematically it is possible to find out which nodes of a deep neural network were activated, but we don’t know what there neurons were supposed to model and what these layers of neurons were doing collectively. So we fail to interpret the results.
|
When should you use ML and DL?
In the table below, you can see the comparative analysis of Machine Learning and Deep Learning.
| |
Machine learning
|
Deep learning
|
|
What about Training dataset?
|
Small
|
Large
|
|
It Chooses features?
|
Yes
|
No
|
|
Number of algorithms
|
Many
|
Few
|
|
Training time
|
Short
|
Long
|
Conclusion:
Thus, as shown in below figure the main difference between machine learning and deep learning is in its working process.
In machine learning you have to identify the features or label the pictures for network to identify while in deep learning the layers of networks identifies the features automatically and then in next layers it finds more complex features and then finally it recognizes the picture. So, in deep learning labelling of data is not needed.
References:
https://towardsdatascience.com/machine-learning-vs-deep-learning-62137a1c9842
https://www.analyticsvidhya.com/blog/2017/04/comparison-between-deep-learning-machine-learning/
https://hackernoon.com/deep-learning-vs-machine-learning-a-simple-explanation-47405b3eef08
https://en.wikipedia.org/wiki/Deep_learning
https://en.wikipedia.org/wiki/Machine_learning
You may also like:
