Personalizations for Associative Learning on Textbooks
Create virtual env
python -m venv .venv
Install packages
pip install --upgrade pip wheel
pip install -r requirements.txt
Generate personalized chapter(s)
python gen.py -c path/to/original_chapter.md -i "<user interest>" -s "<strategy>"
Evaluate/compare personalized chapters
python eval.py -a path/to/final_draft_one.md -b path/to/final_draft_two.md -i "user interest"
- understand student profile
- understand how to best explain CS to this particular student
- personalize structure-based
LLM A: Simulate student judge LLM B: Simulate CS professor judge LLM C: PS
Given a student's bio, extract interest
PS personalize -> LLM A gives feedback & LLM B gives feedback -> PS improve personalization based on both feedbacks (2-step)
"How would you introduce the CS concept of static variables vs instance variables to a student interested in astrophysics?" "Introduce the CS concept of static variables vs instance variables to a student interested in astrophysics."
Can we just have both? So one structurally similar one and another that is more analogy-driven?
have an relation section at the end of the final draft that is just one/two-paragraph analogies per concept -- analogy paras dont need code