What is the MACHINE LEARNING in AI

Machine learning is one of the parts of artificial intelligence or we can say it is an application of AI which can provide systems the ability to automatically run and improve by some experiences. In other words, we can say that machine learning makes the computer system into self-learning mode without any explicit programming. Arthur Samuel invented the machine learning programs in IBM’S Poughkeepsie laboratory in 1959. The first program that created by Samuel was play checkers. There are mainly three types of machine learning such as supervised learning, unsupervised learning and reinforcement learning.

1.Supervised learning in Machine Learning:

  1. Supervised learning is the types of machine learning where both inputs and outputs are already available and the machine predicts the output.
  2. In this learning, a model is created on the basis of training data or level data basis. A user can create a new model and check whether it gives valid output or not. Navies Bayes algorithm is basically based on supervised learning.
  3. This learning can be used in risk assessment, image classification, fraud detection, spam filtering, etc.
  4. Its basically of two types i.e. regression and classification.
    • regression:
      1. regression algorithm is used if there is a relationship between the input variable and the output variable.
      2. It is used in prediction like weather forecasting, market trends, etc.
      3. The regression algorithm which comes under this learning is linear regression, regression trees, nonlinear regression, Bayesian linear regression, polynomial regression.
    • Classification:
      1. Classification is used when the output variable is categorical, which means there are two classes like yes-no, male-female, true-false, etc.
      2. The classification algorithm which comes under the learning is the random forest, decision trees, logistic regression, support vector machines, etc.

Advantages of supervised learning:

  1. By the help of this learning the model can be predict the output on the basis of prior experiences.
  2. In this learning we have an exact idea about the classes of objects.
  3. It helps to solve real world problems such as fraud detection, spam filtering etc.

Disadvantages of supervised learning:

  1. It is not suitable for handling complex tasks.
  2. It can not predict the correct output if the test data is different from training dataset.

2.Unsupervised learning in Machine Learning:

  • In unsupervised learning there are no supervisions, in other words, the algorithm in learning is trained using the data that is unlabeled.
  • The clusters analysis method is a basic common method in the unsupervised learning method.it helps to find all kinds of unknown patterns in data.
  • Its basically two types I.e. clustering and association.
  • Clustering:
  • The method of dividing the objects into clusters which are similar between them and are dissimilar to the objects belonging to other clusters.
  • Association:
  • Discovering the probability of the co-occurrence of the items in a collection.

 The most commonly used unsupervised learning algorithms are k-means clustering, hierarchical clustering, apriori algorithm etc.

Advantages of unsupervised learning:

  1. It easily identifies trends and patterns.
  2. It’s good at handling data that are multi-dimensional and multivariate.

Disadvantages of unsupervised learning:

  1. It requires massive data sets.
  2. It takes enough time to learn algorithms.
  3. High error susceptibility.

3.Reinforcement learning in Machine Learning:

It is basically coming from supervised learning. It’s based on the reward or policy method. Suppose there is an agent and perform an action in an environment and got some reward or penalty on the basis of actions. The state becomes the change and it makes policy and uses the policy in a different way.

The applications used in this learning are robotics for industrial automation and business strategy planning, aircraft control, robot motion control etc.

Advantages of reinforcement learning:

  1. It helps to find situation needs an action.
  2. It provides a learning agent with a reward function.

Disadvantages of reinforcement learning:

  1. It is very time consuming.

Advantages of machine learning:

  1. Automation of everything.
  2. Wide range of applications.
  3. Best for education.

Disadvantages of machine learning:

  1. Possibility of high error.
  2. Data acquisition.

Read Introduction to Artificial Intelligence by me.

Once There was new about AI and Sanskrit connection. Recently, Elon Musk announced First Brain-Machine Interphase

Author: Abhisek Sahoo.

Subscribe to our channel Physics Only

Uploaded by Physics Only so Follow PhysicsOnly for all updates.

Spread the love

Bhautik Kapadiya

Bhautik kapadiya is Scientific Blogger, Scientific Influencer, Web developer, Aspiring Engineer, and Scientific Vlogger. He is Founder of Physicsonly.com. He is Creator of Community of 300k+ Scientific Minds across the Social Media Platforms.