Machine Learning

Machine Learning (ML) is the term used to describe the science of algorithms that programs and applications use to analyze data and create probabilities of outcome to inform the program or application on how to perform a task. Most models rely on input-output pairs to direct the desired results.  A common misconception is that machine learning is separate from artificial intelligence. Furthermore, machine learning is a common implementation of Artificial Intelligence (AI) today and seen as a subset of artificial intelligence.

Machine learning is a technique where a program is trained on a set of data, and the program identifies or learns characteristics of the data that are particularly salient and fit-for-purpose. As the program continues to be used after the initial training, interactions with the program will result in additional training. ML uses statistical techniques to perform the so-called learning and ML algorithms can be used in a large number of fields, including economics, agriculture, brain-machine interfaces, speech recognition, telecommunications, and many other fields.

There are three main types of machine learning that you train the program to learn how to learn, namely:

• Supervised learning: The machine learning task of learning a function that maps an input to an output based on example input-output pairs. It infers a function from labeled training data consisting of a set of training examples. In supervised learning, each example is a pair consisting of an input object and a desired output value. A supervised learning algorithm analyzes the training data and produces an inferred function that can be used for mapping new examples. An optimal scenario will allow for the algorithm to correctly determine the class labels for unseen instances. This requires the learning algorithm to generalize from the training data to unseen situations in a “reasonable” way.
• Unsupervised learning: A type of self-organized learning algorithm that helps find previously unknown patterns in data sets without pre-existing models. And it is also known as self-organization and allows modeling probability densities of given inputs.
• Semi-supervised learning: A hybridization of supervised and unsupervised learning techniques
• Reinforcement learning: An area of machine learning concerned with how software agents ought to take actions in an environment so as to maximize some aspect of cumulative reward. Moreover, it differs from supervised learning in that labelled input-output pairs needn’t be presented, and sub-optimal actions needn’t be explicitly corrected. Instead, the focus is on finding a balance between exploration of previously unknown territory and understanding of current knowledge.

There are mainly three components of machine learning, namely Representation, Evaluation, and Optimization.

• Representation: focuses on determining what the model looks like and how the knowledge is represented.
• Evaluation: Concerned with how good models are differentiated and how candidate programs are evaluated.
• Optimization: Emphasizes on the process of determining good models and how the programs are generated.

In addition to the elements explained earlier, machine learning concept also contains Data Set, Algorithms, Models Feature Extraction, and training components.