In this post I'm taking a look at Lingvist, a spaced repetition platform. Lingvist uses neural networks and machine learning to offer some of the most innovative language learning software on the planet.
Almost everyone has tried to use flashcards to learn things with.
Whether it's the old-fashioned way of writing the question on one side of a card, and the answer on the other – or online where it's all happening automatically.
And the crazy part is they work ridiculously well.
Spaced repetition works by repeating the information at tactically increasing delays, and hence telling the brain that this piece of information is important.
So for language learning the ideal time to see a word you're trying to learn might look something like this:
20 minutes, 1 hour, 4 hours, 1 day, 3 days, 7 days, 2 weeks, 6 weeks, 12 weeks…etc etc etc..
The more reviews you do of a piece of information, the longer you can wait until reviewing it again.
This reduces the risk of forgetting the information.
Sounds easy right? Just space the words out and you have a winner.
Well, yes and no.
In this review I'm taking a deep look at how Lingvist is trying to accomplish this.
Table of Contents:
- Table of Contents:
- The challenge of spaced repetition
- Using machine learning for optimised spaced repetition
- The mission and foundation of Lingvist
- Lingvist Features
- Lingvist Unlimited
- Lingvist Free
- The main interface
- Course Wizard
- Lingvist Challenges
- Which languages can you learn with Lingvist?
- What does Lingvist cost?
- Lingvist Review Conclusion
- Lingvist Alternatives
The challenge of spaced repetition
The biggest problem with implementing the theory of spaced repetition is that we're all different.
We all have different learning backgrounds, histories, cultural awareness, and even personalities.
So, how could you possibly build a system that accounts for everyone whilst still maintaining the highest accuracy of reviews?
You see, in order for a review of a piece of information (like a word in a foreign language) to be the most effective, it should come JUST before you were about to forget it.
Come too late and you've forgotten it, and come too early and you might score poorly on future revisions.
Using machine learning for optimised spaced repetition
This is where software like Lingvist comes in.
Instead of setting arbitrary delays between revisions, or even worse, asking the user themselves to identify “if they have learnt a word or not” Lingvist uses an algorithm that optimises everything for your learner profile.
This is visible in how you start out your new profile:
