Logical methods for AI

Study

Learning goals, recall, chatbots, and flashcards.

This part of the site is about studying the course, rather than about logic. It covers what the learning goals ask of you, a bit of advice on testing yourself, and some tools for doing that.

Learning goals

The course as a whole has learning goals, and so does every chapter. They all start with a verb: define, explain, apply, compare, assess. That’s deliberate. We follow Bloom’s taxonomy (opens in a new tab) , which sorts the things you can do with a piece of knowledge into levels, from remembering it to judging with it. The verb says which level a goal is at.

LevelVerbs we useWhat we’d ask you to do
Rememberdefine, stateGive the definition of a valid inference.
Understandexplain, describe, compareExplain the difference between deductive and inductive validity.
Applyapply, use, represent, implementParse a formula; build the truth table; represent a fact in the language.
Analysedistinguish, analyseSay which reasoning pattern an argument uses.
Evaluateassess, evaluateDecide whether an inference is valid, and say why.

The levels build on each other. Our exercises and exams mostly sit at the upper ones: you’re asked to parse something, calculate something, or judge a case, and that always takes the definitions along with it. A goal that says apply is tested by applying.

Active recall

Active recall (opens in a new tab) means answering from memory, with the book closed, before you look anything up. It takes more effort than reading the chapter again, and it works better: retrieving something is itself what makes it stick.

This is the idea behind the exercises, and behind the flashcards below. Attempt a tutorial question before you read its solution, and redo a worked example on an empty sheet of paper. If you’d rather practise definitions, the flashcards are built for that.