Machine Learning as a Concept of Artificial Intelligence

Machine Learning as a Concept of Artificial Intelligence
Student Name
Program Name or Degree Name 
COURSE XXX: Title of Course
Instructor Name
Month XX, 202X Machine Learning as a Concept of Artificial Intelligence Artificial Intelligence (AI) is among the emerging transformative technologies in modern science. AI develops computer systems that rely on intelligence to perform tasks (West & Allen, 2018). Traditionally, the tasks that require intelligence can only be performed by a human being. Examples of tasks that require human intelligence include speech recognition and perception, language translation, and decision-making. At its core, artificial intelligence is a combination of different concepts and theories that empower machines to be intelligent. Among these concepts is Machine Learning. Description Machine Learning is the branch of AI that utilizes algorithms and data to imitate the way humans learn and behave in different or specific situations. In other words, machine learning is based on the idea that computer systems can learn to make decisions from data patterns with minimal intervention by humans. Therefore, the purpose of machine learning is to delegate roles that humans traditionally performed to machines. The advantage of this aspect is the improved efficiency, especially in situations involving repetitive tasks. The key pillars of machine learning are data and algorithms. Data is vital because it provides the patterns from which machines can learn to perform tasks. In essence, machine learning is impossible where there is no data. On the other hand, an algorithm refers to a finite sequence of computer-readable instructions designed, through programming, to solve a class of problems or computations. Algorithms turn data patterns into actionable decisions or actions. Currently, there are thousands of different machine learning algorithms in the world. Across all these algorithms, there are three consistently inherent concepts. These include representation, evaluation, and optimizations. Representation is vital because it provides the basis for data presentation. Neural networks, graphical models, and decisions trees are examples of the representation concept. Evaluation is vital because it provides the basis for measuring the effectiveness of algorithms. Furthermore, machine learning can be broken down into four types: supervised, unsupervised, semi-supervised, and reinforcement learning. Supervised machine learning is the most practical and most favored type of machine learning. A practical example of supervised learning is when an input variable (x) always produces an output variable (y). Supervised learning involves mapping the function Y=f(X) so that when there are different inputs for x, there are automated outputs for the corresponding y. It is supervised learning because the machine learns from the training datasets. In unsupervised learning, algorithms are not guided by existing data or variables. For instance, in the example above, unsupervised learning will provide the input data (X) but will not provide the output data (Y) from which machines can learn. Semi-supervised learning combines techniques from the first two types, while reinforcement learning comes as a reward after a seq...