Explainable Learner-Video Interaction Prediction using Multimodal Large Language Models

This repository contains the preprocessing, model training and analysis scripts for the AIED 2026 submission "Explainable Learner-Video Interaction Prediction using Multimodal Large Language Models".

Structure of this repository

Please find the supplementary material mentioned in the paper in the file supplementary_material.pdf.

You can find the trained model checkpoints in the folder models.

The prompts for the CTML Feature Coding are contained in video_coding/coding.py.

Setup

The code requires Python 3.9 or higher and the dependencies listed in requirements.txt.