Strand 3: IT Systems

3.11Robotics, artificial intelligence and expert systems

Machine learning

external image merging-of-mind-and-machine_1.jpg

Machine learning is an area of research and practice that focuses on designing algorithms (a set of instructions) that allow computers to look at data and react to the data. The goal is to make algorithms (instructions/tasks) that allow computers to analyze data from the outside world and then act upon that data in a way that is intelligent. This data can either be from sensors or from an existing database depending on the usage of the type of ‘learning’ that the computer is going to do. A big focus in machine learning is making computers that can recognise complex patterns and make intelligent decisions.
Machine learning is used in things like monitoring stock markets, search engines, speech recognition, handwriting recognition, etc. Some of the goals of developed machine learning systems it to eliminate the need for humans while other systems have an incorporation of man and machine working together.

One of the problems of machine learning is that much of the world is very complex for computers and so is hard for them to adapt to or learn from world exerinces.
There are different types of algorithms (or areas) for machine learning:

  • Supervised learning: Generates a function that maps inputs to desired outputs. For example, in a classification problem, the learner approximates a function mapping a vector into classes by looking at input-output examples of the function.
  • Unsupervised learning: Models a set of inputs: like clustering
  • Semi-supervised learning - Combines both labeled and unlabeled examples to generate an appropriate function or classifier.
  • Reinforcement learning - Learns how to act given an observation of the world. Every action has some impact in the environment, and the environment provides feedback in the form of rewards that guides the learning algorithm.
  • Transduction - Tries to predict new outputs based on training inputs, training outputs, and test inputs.
  • Learning to learn - Learns its own inductive bias based on previous experience.
  • Pareto-based multi-objective learning - a Pareto-based approach to learning that results in a set of learning models, which typically trade off between performance and complexity.

- Wikipedia