Unit-3

• Computational thinking


  Artificial intelligence
• Artificial intelligence (AI)


  • What AI means
  • How AI is used in real life
  • Some methods used to develop AI
  • The benefits and limitations of AI
• In this unit you will learn

  1. About artificial intelligence (AI)

  2. Some programming methods used to develop artificial intelligence

  3. The advantages and disadvantages of different AI techniques
• This unit uses a case study
• You will look at the work of a radio operator on an Antarctic base
• He receives signals which are displayed as text messages
• He must check the signals to see if they are real human messages or bad signals caused by interference
• He uses a range of different programming techniques to help spot bad signals
• You will make some programs to help with this task
• Example programs with 'good' and 'bad' signals are provided for you to check
• In this activity you will look at an example of image recognition
• There is a grid shape with numbered rows and columns
• One student creates a secret design by shading in some of the cells on a small sheet
• Other students try to guess the design by calling out cell positions on the large grid
• Deep Dream Generator is a website that uses visual images to explore AI
• It provides tools to upload photos and use AI techniques to change them into different colours and styles
• Heuristics


  A rule that helps you make a quick decision, like a guess or rough estimate based on careful thinking about the problem
• Reinforcement


  A type of machine learning where an agent learns by interacting with its environment and receiving rewards or punishments
• Machine learning


  A type of artificial intelligence that allows systems to learn and improve from experience without being explicitly programmed
• Expert systems


  Computer programs that emulate the decision-making ability of a human expert
• Artificial intelligence


  Building computers that can solve problems using human-like judgement
• Decision tree


  A tree-like model of decisions and their possible consequences, used in machine learning and data mining
• Antarctica is a large continent at the far south of the globe, where scientists live and receive messages by radio
• Taz is a radio operator who must check each signal and only pass on good signals, deleting bad ones
• The program provides a series of short example signals, some good and some bad
• The signals are stored in a list called 'signals'
• The program will traverse the list of signals
• Python offers a shortcut to count through each item in the list using a for loop
• The program will get input from the user to determine if each signal is good or bad
• The good signals will be copied to a new list called 'good_signals'
• Checking messages in human language is a problem that is hard for computers, as it does not have clear steps to a solution
• Artificial intelligence (AI) means building computers that can solve problems using human-like judgement
• This unit will introduce methods like heuristics, expert systems, decision trees, and machine learning
• A heuristic is a rule that helps you make a quick decision, like a guess or rough estimate based on careful thinking about the problem
• Heuristics are not always completely accurate, but they provide a quick way to simplify difficult problems
• Examples of heuristics


  • If you see or smell smoke, there might be a fire
  • Food that is not a normal colour is probably bad to eat
  • A broken ladder might not be safe to use
• The heuristic used in this lesson is: Signals with fewer than three characters are bad
• The heuristic program will check the length of each signal and only add signals longer than 2 characters to the 'good_signals' list
• The heuristic program runs quickly without any input from the user
• The heuristic has simplified the task by removing many bad signals, but it has not completely solved the problem as some bad signals are still in the list