AI is nothing but the capability of machine to imitate intelligent human behavior. AI is achieved by mimicking a human brain by understanding how it thinks, how it learns and works while solving any problem. AI includes various areas of specification:
-Game playing
-Expert System e,g number recognition from the number plate
of speedy car
-Natural Language Processing
-Neural Networks
-Robotics etc
Artificial Intelligence is huge term, it includes Machine Learning and Deep Learning. Basically it is the super-set of ML and DL.
Machine learning is a type of AI which provides computers with the ability to learn without being explicitly programmed. Here we do not need to define all the steps of conditions like any other programming application. however we have to train the machine on a data set(training data), viz large enough to create a model which helps to take decisions based on its learning capability.
For Example: We've to train a machine to determine the species of flower; for that we give it flower data set which contains various characteristics along with respective species like; flower name, sepal and petal length, species name, color,number of sepals and petals, life span etc. using this data set machine will create a model which can be used to identify the flower. Next time we pass some characteristics of flower, machine will display the name of the flower by analyzing these characters.
Although it sounds good and easy to implement but it has a limitation. ML is not capable of handling high dimension data where it has to deal with bulk of input and output data, processing such data becomes complex and resources get exhausted. This is termed as Curse Of Dimensionality. So the crucial problems with huge data cannot be solved with ML.
Another big challenge for traditional ML is Feature Extraction. For Example object recognition, handwriting recognition. These tasks cannot be accomplished by ML, clearly we needed a better approach and here came the concept of Deep Learning.
Deep learning is capable of handling high dimensional data. It is also efficient in focusing on the right features on its own that is feature extraction. Deep learning is implemented through Neural Network which is an Artificial Neuron also called as Perceptron and motivation behind the neural network is Biological Neuron. Deep learning attempts to re-engineer a human brain and studies basic units of brain called Brain cells/Neuron.
DL is very vast and cannot be narrated in one paragraph so I am going to write on Deep learning with more exhaustive details in my coming blogs
-Game playing
-Expert System e,g number recognition from the number plate
of speedy car
-Natural Language Processing
-Neural Networks
-Robotics etc
Artificial Intelligence is huge term, it includes Machine Learning and Deep Learning. Basically it is the super-set of ML and DL.
Machine learning is a type of AI which provides computers with the ability to learn without being explicitly programmed. Here we do not need to define all the steps of conditions like any other programming application. however we have to train the machine on a data set(training data), viz large enough to create a model which helps to take decisions based on its learning capability.
For Example: We've to train a machine to determine the species of flower; for that we give it flower data set which contains various characteristics along with respective species like; flower name, sepal and petal length, species name, color,number of sepals and petals, life span etc. using this data set machine will create a model which can be used to identify the flower. Next time we pass some characteristics of flower, machine will display the name of the flower by analyzing these characters.
Although it sounds good and easy to implement but it has a limitation. ML is not capable of handling high dimension data where it has to deal with bulk of input and output data, processing such data becomes complex and resources get exhausted. This is termed as Curse Of Dimensionality. So the crucial problems with huge data cannot be solved with ML.
Another big challenge for traditional ML is Feature Extraction. For Example object recognition, handwriting recognition. These tasks cannot be accomplished by ML, clearly we needed a better approach and here came the concept of Deep Learning.
Deep learning is capable of handling high dimensional data. It is also efficient in focusing on the right features on its own that is feature extraction. Deep learning is implemented through Neural Network which is an Artificial Neuron also called as Perceptron and motivation behind the neural network is Biological Neuron. Deep learning attempts to re-engineer a human brain and studies basic units of brain called Brain cells/Neuron.
DL is very vast and cannot be narrated in one paragraph so I am going to write on Deep learning with more exhaustive details in my coming blogs
No comments:
Post a Comment