Enhancing Small Language Models for Power Outage Report Generation via Minimum Risk Training

2026-09-22

AI Summary

该研究通过最小风险训练优化序列级评估指标，提高小规模语言模型在电力中断报告生成中的准确性，将异构报告转换为符合CIM IEC 61968-3标准的XML格式。

Abstract

Minimum Risk Training (MRT) enables neural machine translation models to directly optimize sequence-level evaluation metrics instead of relying only on token- level maximum-likelihood objectives Shen et al. [2016]. Although introduced a decade ago, recent work shows renewed potential for risk-based optimization in modern language models Yang et al. [2024], Jinnai et al. [2025]. We apply MRT to power outage report generation for the Outage Data Initiative Nationwide (ODIN), transforming heterogeneous reports into standardized XML compliant with CIM IEC 61968-3. Our MRT approach improves Qwen2.5-7B-Instruct overall accuracy from 16.20% to 68.95%, demonstrating the effectiveness of sequence- level optimization for domain-specific structured generation

Problem

• Minimum Risk Training
• Power Outage Report Generation
• Sequence-level Optimization

Innovation

• Minimum Risk Training
• Sequence-level Optimization
• Power Outage Report Generation