Know automatically if the user gets it right. (Currently, the user must score himself by clicking the “Right” or “Wrong” button.

Use the microphone, compare user's speech input, and decide whether it's right or wrong.

Let the user type in the answer, and do a text string compare to see if it is correct.

Reading, writing, listening, speaking.

User can input by typing, drawing, or speaking.

1. typing - string compare
2. drawing - AI pattern matching of pen strokes
3. speaking - use microphone and Web Speech API to convert input to text, then to a string compare

The algorithm for manipulating the stacks of flash cards is implemented in coach.js.

The algorithm is based on parameters.

At present, we set the parameters manually.

We should be able to use the autoplay feature, repetitively altering the parameters striving for optimal success in mastering the cards.

We should also introduce parameters describing the human.

• number of cards in working set
• average number of attempts to learn a new cards

Set up a meta-learning AI who repetitively sets up decks and parameter sets and runs autoplay, saving the result of each run, gradually improving the parameter set.

class Train

• set parameters
• run with autoplay
• record result: number of attempts to master all cards
• loop
  • adjust parameters
  • run with autoplay, max 10000 runaway escape
  • record result

adjust parameters

• a weight multiplied times each parameter
• vary the weight up, see the effect on result
• vary the weight down, see the effect on result

A never-ending stream of cards is exhausting. Break it into sets. Give the user a completion, celebration, and rest before the next set.

Music. Explosions.

Note the user's speed and accuracy, compared to his average, and guess at his emotional state, or at least his levels of energy and focus.