Montessori-Instruct: Generate Influential Training Data Tailored for Student Learning

TL;DR

Montessori-Instruct is a novel data synthesis framework that tailors synthetic training data to enhance student language model learning by leveraging local data influence and direct preference optimization, significantly improving performance.

Contribution

It introduces a new method for generating student-specific synthetic data using local influence and DPO, outperforming standard methods and stronger teachers.

Findings

Outperforms standard synthesis methods by 18.35% and 46.24% on benchmarks.

Surpasses data from a stronger teacher model, GPT-4o.

Demonstrates robustness across different student models.

Abstract

Synthetic data has been widely used to train large language models, but their generative nature inevitably introduces noisy, non-informative, and misleading learning signals. In this paper, we propose Montessori-Instruct, a novel data synthesis framework that tailors the data synthesis ability of the teacher language model toward the student language model's learning process. Specifically, we utilize local data influence of synthetic training data points on students to characterize students' learning preferences. Then, we train the teacher model with Direct Preference Optimization (DPO) to generate synthetic data tailored toward student learning preferences. Experiments with Llama3-8B-Instruct (teacher) and Llama3-8B (student) on Alpaca Eval and MT-Bench demonstrate that Montessori-Instruct significantly outperforms standard synthesis methods by 18.35\% and 46.24\% relatively. Our method also beats data synthesized by a stronger teacher model, GPT-4o. Further analysis confirms the benefits of teacher's learning to generate more influential training data in the student's improved learning, the advantages of local data influence in accurately measuring student preferences, and the robustness of Montessori-Instruct across different student models. Our code and data are open-sourced at https://github.com/cxcscmu/Montessori-Instruct.