Skip to main content

Project - Kyoto, baby

AI-assisted learning for neural networks.

Team Members:

    Carolina Esna Stergios

    Core idea: 

    Learning goals: 

    Students will be able to implement neural network concepts. 

    Research question:

    • How AI-assisted interactive visualisation of neural networks improve student learning performance?

    • Would learning platform with AI feedback and instructions enhance stduent learning performance?

    Task Analysis:

    https://playground.tensorflow.org/ clone with specific targets.

    • Experimental Group (A): Users get AI assistance for the first 10 examples (clipy style), then they have 5 test cases without AI.

    • Control Group (B): Users get 10 test examples they need to optimise for, with generic rules of how different parameters affect performance, then 5 test cases

    Recruitment plan: 

    • Master's student (Michalis's)

    • Master EPFL student (Mauro's)  ---> Isna need to asked first

    We record (everything, but we focus on):

    • Performance on the 5 test questions (complexity vs performance) and speed

    • Engagement between clicks

    • If Michalis approves -> performance on the NN assignment (post-intervention performance)

    • Self-assessment survey: AI literacy, Intrinsic motivation