By Sam Gorman and Kanyes Thaker
August 18, 2021
Ferret is a Chrome extension that uses NLP to generate helpful recall-based questions to reinforce key concepts for almost anything you read on the internet.
An example of Ferret in action, helping users learn and remember what they read on the internet. In this case, we're using Ferret to learn React.JS documentation.
Let's begin with a hypothetical situation that may hit home to some readers.
You're doing your best to learn a difficult concept that interests you. You read material online to do so: perhaps class notes, documentation, an article, an online textbook, a forum, or a blog post. After some period of research, you find a set of information that's useful for you. After reading, you finally feel as if you've gotten a grasp on this material. But a day or two later, when asked to explain what you read, you come up blank or only remember the most superficial of details. What happened?
Even in the best of times, human memory is lossy. And learning an intricate topic is often more like sailing a leaky ship than crossing a bridge. Psychologist Hermann Ebbinghaus found as early as 1885 that human memory may be modeled by a forgetting curve, where information is quickly lost over time when there is no attempt to retain it.
The products we use to learn everyday generally haven't helped us make any attempts to retain our knowledge. Classical services like Quizlet made a name for themselves helping students memorize vocabulary for tests. And the most dedicated configure personal systems with tools such as Anki in attempts to augment their memory. But these approaches miss a fundamental truth:
Recent approaches to augmenting memory in learning have shown progress over more classical approaches such as Quizlet and Anki. Researchers Andy Matuschak and Michael Nielsen introduced the Mmnemonic medium in 2019, and demonstrated its potential in Quantum Country, an experimental essay that blends technical content and inline spaced-repetition flashcards. Open source approaches such as Orbit provide tools for authors to implement their own inline spaced-repetition flashcards such as those seen in Quantum Country.