Hacker News Reader

Comment by henry_pulver | Hacker News Reader

What actually is incremental learning?

The 'What is Incremental Learning?' section seems deliberately poorly written.

    Incremental learning is the fastest and the most comprehensive way of learning available to students at the moment of writing (2013).

    Incremental learning is a consolidation of computer-based techniques that accelerate and optimize the process of learning from all conceivable material available in electronic form, and not only.

    Currently, SuperMemo is the only software that implements incremental learning. In SuperMemo, the student feeds the program with all forms of learning material and/or data (texts, pictures, videos, sounds, etc.). Those learning materials are then gradually converted into durable knowledge that can last a lifetime.

View on HN

I’m not sure, but learning is certainly more than remembering things. Years ago I spent a lot of time typing CS things into Anki and practicing their recall. Eventually I realised that I didn’t really have a good understanding of half the things I was memorising.

Retrieval practice is one of the most effective ways to learn and retain new information. Research shows that it's more effective than summarization, elaborate interrogation, self-explanation, highlighting, imagery use for text learning, rereading and interleaved practice [0].

Not only that, retrieval practice has shown not only to boost retention but meaningful learning [1] as well. For example, in one study, students who studied a given text solely by practicing retrieval performed significantly better on creating concept maps than students who studied by creating concept maps.

Retrieval practice is not the only superweapon Anki entertains in its arsenal, it also employs distributed practice, which is another very powerful tool and finding in cognitive research with plenty of evidence supporting it.

[0]: https://sci-hub.hkvisa.net/10.1177/1529100612453266

[1]: https://sci-hub.hkvisa.net/10.1126/science.1199327

Yes, you’re not supposed to blindly memorize things. The method is to learn and understand the material first and then practice recall to retain that understanding.

It's a system for automating learning.

You have thirty books. Instead of taking a month for each book, getting value sequentially, you read all thirty books at the same time, with the software splitting it up for you into tiny pieces, so you can soak them all in and get lots of knowledge simultaneously, rather than the sort of limited, controlled drip of knowledge from an individual work.

It's surprisingly useful for things like language learning, where you don't really start with much understanding but books intend for your ability to change while reading; that said, the author uses it for everything he consumes, even email.

It's pretty interesting, if you're of the belief that knowledge compounds, or if you just don't really have the patience or attention span for individual works. Certainly isn't for everyone, and if you aren't willing to make your own tooling, the existing tooling is pretty bad.

This made me wonder, which languages would be best learned together. Probably a pairing like Italian, French, Spanish would work well, as they're all Latin languages.

Or perhaps it would be better to pair some languages which have defining differences.

I think similar languages such as the ones you mention would be great; as a native spanish speaker I continue to find joyful coincidences and contrasts that help even better understanding my native tongue. For example, neither guardare (italian), nor regarder (french), have a direct reflection in spanish (guardar in spanish is to keep, or to look after, but not literally to look, which is mirar instead), but it is what a guardia does.

I've found learning very similar languages simultaneously is tricky because cognates conflict with each other. Is it lemon or límon? It might be better to get good enough at one that you can intuitively distinguish it from the other, without pausing to reflect.

Reply on news.ycombinator.com