Make your own Lesson Plan for Music Blocks

[walterbender]

Ticket Contents

Description

We would like to deploy generative AI in support of converting Music Blocks projects into lesson plans. An LLM would be used to convert the musical and computational ideas expressed in a project to auto-generate the first draft of lesson plan.

Some examples of lesson plans can be found here and here.

Specifically, we would be working toward accomplishing the following:

Create scripts that can parse Music Blocks JSON data into data analyzable by LLM
Input existing lesson plan data into LLM
Tune LLM to create useful lesson plans

Goals & Mid-Point Milestone

Goals

• [ ] [Lesson Template]
• [ ] [LLM backend]
• [ ] [Block extraction]
• [ ] [Frontend interface]
• [ ] [Goals Achieved By Mid-point Milestone]
• [ ] [Template]
• [ ] [Backend]

Setup/Installation

No response

Expected Outcome

The user will be able to engage with an LLM around a lesson idea, expressed both in Music Blocks code and in a dialog. The output of the engagement is a structured document which contains an editable first-draft lesson plan.

Acceptance Criteria

The basic lesson plan structure must be reflected in output. The UX should include both project code and a dialog with the instructor developing the lesson plan.

Implementation Details

Presumably the LLM will be hosted in a docker image on the Sugar servers. The lesson plan generation can be stand-alone, with access to projects stored on the Planet.

Mockups/Wireframes

No response

Product Name

Music Blocks

Organisation Name

Sugar Labs

Domain

⁠Education

Tech Skills Needed

Artificial Intelligence, ASP.NET, Docker, JavaScript, Python

Mentor(s)

@pikurasa @walterbender

Category

Backend, Documentation, Frontend, Research, AI