few-shot-classification

Proposes a labeled-data-free meta-learning method that generates tasks by assigning soft labels from pre-trained models to unlabeled data, avoiding computationally expensive model inversion. Achieves up to 104x speedup and 8.4-36.4% accuracy improvements over state-of-the-art DFML methods.

The paper presents a pipeline that maps student questions from a conversational AI teaching assistant to curriculum topics using a few-shot text classifier and a GPT-4-extracted prerequisite knowledge graph, achieving 80% accuracy on 1,340 question events and correlating with self-reported difficulty.