Open-source AI concept map: feedback on prerequisites and learning order

AI Discussions
1

Hi, I’m Jingwen, the maintainer of AI Knowledge Map. I built this free, open-source resource to help beginners see how AI concepts connect before choosing what to study next. It includes 130 concepts, 416 typed relationships and nine suggested learning stages, in English and Simplified Chinese. Reading does not require an account.

I’d appreciate feedback on a specific curriculum-design problem: distinguishing a real prerequisite from a merely related concept. For example, for a beginner’s first explanation of retrieval-augmented generation, should embeddings and similarity search be required first, or introduced alongside retrieval? How much language-model background is enough before that explanation becomes useful?

If you try the map, please point out one dependency that seems missing or unnecessary, or one place where the suggested learning order would confuse a newcomer. Concrete counterexamples would help me improve it.

Website: https://ai-knowledge-map.com/
Source: GitHub - Little-FishK/AI-knowledge-map: Free bilingual AI learning map for beginners: 130 concepts, a 9-stage roadmap, concept explainers, and progress tracking. · GitHub

Disclosure: this is my own project. The content is still being reviewed, and I have not completed a formal learning-outcomes study. I’m sharing it for technical and educational feedback, and would be happy to discuss suggestions here.

2

As someone transitioning into AI with no CS background, this is a very cool tool and definitely has a lot of utility to someone like me. Aside from the tool itself, I think it is very well laid out and I didn’t have any trouble or confusion navigating the site. I also like the required models to explain the high level concepts before accessing the map (unless i didn’t need to do that, but it was very natural to go through the modules first.

I do have a fairly good understanding of these beginner concepts after working with these kinds of projects for about a year, but it was still useful to skim through. I think the interactive map is what is most useful to someone like me. It does not feel like information overload, and the UI incentivizes exploration.

Amazing job. Excited to see future iterations. Sorry this is more a review than constructive criticism. But I hope it’s affirming of your idea.