Optimizing Instruction Synthesis: Effective Exploration of Evolutionary Space with Tree Search
This addresses the labor-intensive challenge of creating high-quality instruction data for language model alignment, offering a scalable solution with measurable gains, though it is incremental as it builds on existing iterative enhancement methods.
The paper tackles the problem of low-quality and uncertain instruction data synthesis for aligning language models by introducing IDEA-MCTS, a framework that uses tree search to guide instruction evolution, resulting in improved average evaluation scores from 2.19 to 3.81 and a 5% accuracy boost in low-resource settings.
Instruction tuning is a crucial technique for aligning language models with humans' actual goals in the real world. Extensive research has highlighted the quality of instruction data is essential for the success of this alignment. However, creating high-quality data manually is labor-intensive and time-consuming, which leads researchers to explore using LLMs to synthesize data. Recent studies have focused on using a stronger LLM to iteratively enhance existing instruction data, showing promising results. Nevertheless, previous work often lacks control over the evolution direction, resulting in high uncertainty in the data synthesis process and low-quality instructions. In this paper, we introduce a general and scalable framework, IDEA-MCTS (Instruction Data Enhancement using Monte Carlo Tree Search), a scalable framework for efficiently synthesizing instructions. With tree search and evaluation models, it can efficiently guide each instruction to evolve into a high-quality form, aiding in instruction fine-tuning. Experimental results show that IDEA-MCTS significantly enhances the seed instruction data, raising the average evaluation scores of quality, diversity, and complexity from 2.19 to 3.81. Furthermore, in open-domain benchmarks, experimental results show that IDEA-MCTS improves the accuracy of real-world instruction-following skills in LLMs by an average of 5\% in low-resource settings.