Deep Learning vs Machine Learning
What is Deep Learning?
Deep Learning involves a computation model that is composed of multiple processing layers, which learns representations of raw data by multiple levels of abstraction.
Now let’s dive deeper into the details:-
Deep Learning is a subset of artificial intelligence that imitates the working of the human brain (in the form of a neural network) in processing data and creating patterns for use in decision making.
In deep learning, we don’t need to explicitly program everything. Deep learning learns from a vast amount of unstructured data that would normally take humans decades to understand and process. Deep learning models can solve complex problems.
In simple words we can say that deep learning is the type of machine learning algorithm that divides the input into layers and it classifies, generates, and predicts data with much greater efficiency.
One very remarkable fact about deep learning can be understood from the following quote given by Ilya Sutskever :
“ If you have a large big dataset and you train a very big neural network, then success is guaranteed! ”
Deep Learning learns through examples, like a human baby learning from its surrounding world.
Applications of Deep Learning :
- Self-driving car
- Automatic text generation
- Virtual Assistant
- Natural language processing
- Pattern Recognition
- Computer Vision and much more
Types of networks used to develop Deep Learning models :
- Convolution neural network:
It is a type of neural network containing filters that are not hardcoded, rather learned during the training process in order to extract relevant features from the input distribution.
2. Recurrent neural network:
It is a type of neural network that contains loops, allowing information to be stored within the network. In short, Recurrent Neural Networks use their reasoning from previous experiences to inform the upcoming events.
Advantages of Deep Learning :
- Maximum utilization of unstructured data
- Elimination of the need for feature engineering
- Ability to deliver high-quality results
- Elimination of unnecessary costs
- Elimination of the need for data labeling
Disadvantages :
- Require a large amount of data
- Expensive to implement its model
Following is the link to some slides to get a better understanding of neural networks:
Ml and DL used web apps:
This is the demonstration of flappy bird game in which bird learns to play the game using reinforcement learning(a part of deep learning) and trains to get higher score in each round.
2. https://quickdraw.withgoogle.com/shared/vlNg0sr2MhFm
This is a fun web app in which we trained a large dataset of doodle images with the help of neural networks and we have asked to draw a particular doodle and our model predicts it is correct or not.
Deep Learning over Machine Learning
- Why I can’t get the efficiency of 90%
- How to fit my model for unlabelled data
To answer these doubts, deep learning comes into the picture !!
In order to find out when and where deep learning proves to be a better choice over machine learning, let’s first understand a bit about machine learning.
Machine Learning is a set of algorithms that parse data, learn from them, and then apply what they’ve learned to make intelligent decisions. Example: Facebook recognizes your friend’s face in a digital photo
Whereas Deep Learning is a neural network which studies data, classifies it, and develops patterns to make predictions.
What Deep learning can do but machine learning can’t?
- It can develop fake new data from existing real data
- It can develop games by reinforcement learning
- It generates a model of greater efficiency
The idea behind machine learning is that the machine can learn without human intervention. The machine needs to find a way to learn how to solve a task given the data.
Deep learning is a breakthrough in the field of artificial intelligence. When there is enough data to train on, deep learning achieves impressive results, especially for image recognition and text translation. The main reason is the feature extraction is done automatically in the different layers of the network.
Decision making procedure in ML and DL:
Machine learning uses algorithms to parse data, learn from that data, and make informed decisions based on what it has learned.
Deep learning structures algorithms in layers to create an “ artificial neural network ” that can learn and make intelligent decisions on its own.
Have a look at the following blog to know more about how deep learning is used in practice:
Two simple scenarios to show where ML, DL is used
Scenario-1 :
You are given the following dataset (showing only 5 rows of it, suppose the dataset has 200 rows )
This dataset is used to predict whether a person will have a fever or not.
Training data = Given table data
Testing data= (Sanya,yes,yes,yes)
Output Predicted = yes
Did this problem be solved by machine learning or deep learning? So to decide this, notice carefully a few things :
- The given dataset is quite small.
- After studying the table, we note that features like “ consume cold drinks ”, “set AC at low temp ”, “has cold” have a great influence on whether the person would have a fever or not while rest features don’t seem to have any relation with the target variable.
- Since the dataset is small and we have less no of the important features mentioned, Machine learning can be employed in this case.
Scenario-2 :
You are given the following dataset ( showing only 5 rows and 11 features, suppose the dataset has 2,00,000 rows and 500 features)
This dataset is used to predict whether a person will have a fever or not.
Training data = Given table data
Testing data= (Sanya,yes,yes,yes,no,45,no,no,yes,12,day,…..)
Output Predicted = no
Should this problem be solved by using machine learning or by employing deep learning? To decide this, a few things can be noted:
- You can see this is quite a large dataset as compared to the previous one.
- Also, it’s difficult to decide which feature is more contributing to the target variable.
- Therefore deep learning is used in such cases.
Hope this blog could help you get a better understanding of deep learning, its importance, and its need in this digital era.
Thanks for your time!
