Resource roadmap

[SkalskiP]

I think one of the biggest problems with similar repositories like ours is:

• Too many resources that overwhelm people.
• Lack of direction. People don't know what they have to do first to achieve the goal they have set for themselves.

Therefore, I want to prepare a roadmap to show people which courses are most valuable. It will be a tree of knowledge - root: I know nothing, leaf: I have knowledge of the given topic. I need to refine my idea but it will be something like this:

├─ Foundation
│  ├─ 🧮 Math
│  │     ├─ course link
│  │     ├─ ...
│  │     └─ course link
│  └─ 👩🏻‍💻 Programming
│        ├─ course link
│        ├─ ...
│        └─ course link
.
.
.

[rasbt]

To spin it further:

├─ Foundation
│  ├─ 🧮 Math
│  │     ├─ course link
│  │     ├─ ...
│  │     └─ course link
│  └─ 👩🏻‍💻 Programming
│        ├─ course link
│        ├─ ...
│        └─ course link
|
├─ AI Fundamentals
│  ├─  Machine Learning
│  │     ├─ course link
│  │     ├─ ...
│  │     └─ course link
│  └─ Deep Learning
│  │     ├─ course link
│  │     ├─  ...
│  │     ├─ course link
│  └─ AI
│        ├─ course link
│        ├─ ...
│        └─ course link
|
├─ Natural Language Processing
│       ├─ course link
│       ├─ ...
│       └─ course link
├─ Computer Vision
│       ├─ course link
│       ├─ ...
│       └─ course link
├─ Reinforcement Learning
│      ├─ course link
│      ├─ ...
│      └─ course link

.
.
.

Btw do you think generative AI should be a separate top-level category? It becomes tricky then because it intersects with many NLP and CV resources

[rasbt]

Oh yeah, the formatting doesn't look too bad!

Regarding the level structure ... Maybe NLP & CV should be under "Domain-Specific AI". I would keep RL as a separate top-level one since it's really its own field. Maybe for Generative AI a top-level is ok too because then you can have NLP and CV also as sublevels there.

Multimodal AI is tricky. I haven't watched the course and don't know what it covers. Is it strictly Generative AI-related? (Thinking of Stable Diffusion and other text2image models).

Otherwise, maybe under "Special Topics", where Special Topics is a top-level category.

[SkalskiP]

Oh yeah, the formatting doesn't look too bad!

I will dig deeper tomorrow. Maybe I'll find some better formatting options. I need to put it on pause. I need to do more work :)

If we do not find better formatting options, I'll hide the learning resources roadmap in a clickable section like the one below so that it wouldn't hijack the whole README.md. I'll probably make some PR tomorrow.

👉 learning resources roadmap

│

├─ 🏗️ Foundation

│     ├─ 🧮 Math

│     │     ├─ course link

│     │     ├─ ...

│     │     └─ course link

│     └─ 👩🏻‍💻 Programming

│           ├─ course link

│           ├─ ...

│           └─ course link

│

├─ 🤖 AI Fundamentals

│     ├─  📈 Machine Learning

│     │     ├─ course link

│     │     ├─ ...

│     │     └─ course link

│     ├─ 🧠 Deep Learning

│     │     ├─ course link

│     │     ├─  ...

│     │     └─ course link

│     └─ 🧩 AI

│           ├─ course link

│           ├─ ...

│           └─ course link

│

├─ 💬 Natural Language Processing

│     ├─ course link

│     ├─ ...

│     └─ course link

│

├─ 📷 Computer Vision

│     ├─ Stable Diffusion

│     ├─ ...

│     └─ course link

│

├─ 🎮 Reinforcement Learning

│     ├─ course link

│     ├─ ...

│     └─ course link

│

├─ 🎨 Generative AI

│     ├─ Stable Diffusion

│     ├─ ...

│     └─ course link

.

.

.

Maybe NLP & CV should be under "Domain-Specific AI". I would keep RL as a separate top-level one since it's really its own field. Maybe for Generative AI a top-level is ok too because then you can have NLP and CV also as sublevels there.

I like that.

Otherwise, maybe under "Special Topics", where Special Topics is a top-level category.

I like that too.

[SkalskiP]

The alternative is to drop monospace font and have this. But it looks a bit 💩 in my opinion.

• 🏗️ Foundation
  • 🧮 Math
    • course link
    • ...
    • course link
  • 👩🏻‍💻 Programming
    • course link
    • ...
    • course link
• 🤖 AI Fundamentals
  • 📈 Machine Learning
    • course link
    • ...
    • course link
  • 🧠 Deep Learning
    • course link
    • ...
    • course link
  • 🧩 AI
    • course link
    • ...
    • course link
• 💬 Natural Language Processing
  • course link
  • ...
  • course link
• 📷 Computer Vision
  • course link
  • ...
  • course link
• 🎮 Reinforcement Learning
  • course link
  • ...
  • course link
• 🎨 Generative AI
  • course link
  • ...
  • course link

[rasbt]

I actually don't mind this 😅

[SkalskiP]

Hah... maybe it is just me... If you could pick one:

• monospace font + links, but with bit line space
• font + links, but without monospace font

Which would you choose?

[rasbt]

Don't want to make it even more complicated, but I also like the list style you currently have. So a 3rd option would be have is breaking the current list into smaller lists with different headers.

[SkalskiP]

Haha. This is cool but the problem I see is that when person is interested in CV they ask me what courses are prerequisite for course X. And I’d like to provide them not only with courses divided into groups but also the path they need to follow.