Introduction To Machine Learning


In simple term, Machine learning is the field of study that gives computers the ability to learn without being explicitly programmed. (Arthur Samuel,1959)

And if it is to be described on engineering oriented one:

A computer program is said to learn from experience E with respect to some task T and some performance measure P, if its performance on T, as measured by P, improves with experience E.(Tom Mitchell,1997)

Machine learning is an important component of the growing field of data science. Through the use of statistical methods, algorithms are trained to make classifications or predictions, uncovering key insights within data mining projects. It has become a key technique for solving problems in areas, such as:

  • Image processing and computer vision, for face recognition, motion detection, and object detection
  • Computational finance, for credit scoring and algorithm trading
  • Automotive, aerospace, and manufacturing, for predictive maintenance
  • Natural language processing, for voice recognition applications

Why Use Machine Learning?

For this approach, let begin with an example of speech recognition: say you want to start simple and write a program capable of distinguishing the word “one” and “two”. You might notice that the word “two” starts with a high pitch sound (“T”), so you could hardcore an algorithm that measures high pitch sound intensity and use that to distinguish ones and twos. Obviously, this technique will not be perfectly suitable to millions of people around the world who speak different language and have different accent. So the best solution is to write an algorithm that learns by itself, give many example recordings for the word from around the world.

To summarize, Machine learning is great for:

  • Problems for which require a lot of individual processing data, Machine learning algorithm can learn to find similar pattern and can simplify work and perform better.
  • Complex problems for which there is no good solution at all using a traditional approach:the best machine learning techniques can find solution.
  • Flotation environments: a machine learning system can adapt a new data
  • Getting insights about complex problems and large amounts of data

Types of Machine Learning system

There are so many types of machine learning systems that is useful to classify them in broad categories based on:

  • Supervised, unsupervised, semi-supervised, and reinforcement Learning
  • Online and batch processing
  • Instance based versus model based learning.

For introductory will learn only about Supervised, Unsupervised, semi-supervised, and reinforcement Learning.

Supervised Learning

In supervised machine learning, the training data you feed to the algorithm includes the desired solutions, called labels. It takes a known set of input data and known result to the data and trains a model to generate reasonable predictions for the response to new data.

For example, to predict a target numeric value, such as the price of a car, with a given set of feature such as mileage, age, brand, etc. They have data on previous car and its price according to its feature. So the problem is combining the existing data into a model that can predict the price of a selected car.

Some of the important supervised learning algorithms are

  • Linear Regression
  • Logistic Regression
  • Neural networks
  • support vector machines, etc

Unsupervised Learning

Unsupervised learning uses machine learning algorithms to analyze and group similar unlabeled datasets. These algorithms discover hidden patterns or data groupings without the need for human intervention.

For example, an ISP wants to optimize the locations where they build an optical fiber network, they can use machine learning to estimate the number of clusters of people who need internet with proper bandwidth. So they can use clustering algorithm to design the best way to distribute internet among them and to optimize bandwidth among users.

Some important unsupervised learning algorithms are

  • Clustering
  • Visualization and dimensionality reduction
  • Association rule learning

Semi-supervised learning

Some algorithms can deal with partially labeled training data, usually a lot of unlabeled data and a bit of labeled data. This is called semi-supervised learning. Most semi-supervised algorithms are a combination of both supervised and unsupervised learning.

Photo hosting services, such as Google Photos, are a good example of it. Once you upload all your family photos to the service, it automatically recognizes that the same person A is on photo 5, 8, and 12. This is the unsupervised part of the algorithm. Now all the system need is for you to tell it who these people are. Now just one label per person, and it will be able to name everyone in every photo and categorized it according to person label. If a person search photo with name, it can directly show photos with named person present in it.

Reinforcement machine learning

Reinforcement machine learning is a behavioral machine learning model that is similar to supervised learning, but the algorithm isn’t trained to sample data. This model learns by itself with trial and error. The learning system, called an agent and can observe the environment, select and performs actions, and get rewards in return.

For example, DeepMind Alpha program, which beat world champion Lee Sedol at the game of Go in 2016. It learned its winning policy by analyzing millions of game, and playing with itself. At games against champions, learning was turned off, and it won by just applying policy it has learned.

Read About Web tools:

%d bloggers like this: