Spaced repetition

Anki, a computer program implementing spaced repetition.

Spaced repetition is a learning technique in which increasing intervals of time are used between subsequent reviews, rather than studied frequently for a short time; this exploits the psychological spacing effect. Alternative names include expanding rehearsal, graduated intervals, repetition spacing, repetition scheduling, spaced retrieval and expanded retrieval.

Applications

Practical applications of spaced repetition were first suggested in the book Psychology of Study by Prof. C. A. Mace in 1932.

Pimsleur language courses use spaced repetition techniques, and in the 1970s Sebastian Leitner devised his Leitner system, an all-purpose system based on flashcards.

More recently, spaced repetition has also been implemented into computer-assisted language learning software to gradually adapt the optimum spacing of repetitions to individual needs . A typical optimization criterion used in spaced repetition is the requested level of knowledge retention, i.e. percent of knowledge that is to be remembered.

There are several families of algorithms for scheduling spaced repetition:

• Neural networks based
• Sebastian Leitner system learning machines:
  • 5 stages
  • n stages
• SM-family of algorithms (SuperMemo):
  • SM-0
  • SM-2
  • SM-4
  • SM-5
  • SM-6
  • SM-8
  • SM-11

The precise length of intervals does not have a great impact on algorithm effectiveness, as reflected in data collected by Mnemosyne.

Prominent researchers

• Cecil Alec Mace
• Hermann Ebbinghaus
• Paul Pimsleur
• Sebastian Leitner
• Piotr Woźniak

Software

Most programs are modeled after the original SuperMemo: items to memorize are entered into the program as question-answer pairs; the algorithm schedules pairs, and when a pair due to be reviewed, the question half is displayed on screen, and the user is supposed to attempt to remember the answer or fill in the blank (in the case of a Cloze deletion); when the user has succeeded or failed, she then manually reveals the answer, and then tells the program how easily she recalled the answer (or whether she failed to), at which point the program then displays the next scheduled question.

Further refinements are possible: for example, there could be an option for a pair to be treated both as question-answer and answer-question (useful for vocabulary - one might wish to test defining a given word, and deciding what word matches a given definition); or the program could be web-based; or it could support advanced input formats such as LaTeX.

Implementations:

• Anki
• FullRecall
• Mnemosyne
• SuperMemo
• Winflash

See also

• Study Software
• Testing effect
• Graduated interval recall (System developed by Paul Pimsleur)