We believe that by using natural language processing to model human memory, we can create a better flashcard app. We are testing this idea through KAR³L, an adaptive flashcard system designed to boost your efficiency when studying. We target KAR³L towards trivia and general knowledge enthusiasts, including those studying for Quizbowl or other trivia competitions.
We first launched a version of KAR³L a few years ago and during our first round of testing, we had over 800 sign ups! We’ve since made many improvements to our app and are ready for our second round of testing. To develop the best study plan for each user, KAR³L maintains an individualized student model and uses machine learning to estimate your current skill level. Unlike existing methods, our student model captures the semantic similarity between flashcards to improve student retention predictions.
To help us with this study, we would like you and any interested friends to use KAR³L, which can be accessed at https://karl.qanta.org. KAR³L can also be installed as a desktop, Android, or iOS app! Instructions can be found at https://karl.qanta.org/pwa.
In this round of testing, which lasts until January 2024, we're excited to compare our model’s performance with other leading flashcard systems. Once a day when you log into our app, you'll be asked to participate in a “test mode” session, where a curated set of 10 or 20 flashcards will be presented to you to study. In total, there will be 18 test mode sessions spanning 18 days. Outside of test mode, we encourage users to spend at least 10 minutes a day using our app, but this is not a strict requirement and users can study as little (or as much!) as they would like on any particular day.
KAR³L will continue to be available after January, this won’t be our last phase of testing! To make it more exciting, we're offering an incentive for this phase of tests:
- The first 50 participants who complete a minimum of 12 test mode sessions (lasting 12+ days) will each be awarded a $50 gift card in compensation.
- The rest of the participants who complete 12+ test mode sessions until the end of January 2024 will be entered into a raffle to win one of: two $50 gift cards, five $20 gift cards, or five $10 gift cards
- For participants to receive rewards, they must make a good-faith attempt to study the test mode flashcards. Participants who rapidly flip through the flashcards without providing genuine answers will not receive rewards, up to our own discretion.
If you have any questions, please contact the co-PIs of this research project: Matthew Shu (Yale, [email protected]) and Nishant Balepur (University of Maryland, [email protected]). The PI for this project, Professor Jordan Boyd-Graber (University of Maryland), can be reached at [email protected]. You can also join our Discord at https://discord.gg/PTfEmHd!