[论文] Enhancing Small Language Models for Power Outage Report Generation via...

论文概要

研究领域: ML
作者: Hung Phan, Waqwoya Abebe, Youssef Hussein, Supriya Chinthavali, Dalton Lunga, Ali Jannesari
发布时间: 2026-09-25
arXiv: 2609.27197

中文摘要

最小风险训练（MRT）使神经机器翻译模型能直接优化序列级评估指标，而非仅依赖 token 级最大似然目标（Shen et al. [2016]）。尽管十年前就已提出，近期工作表明基于风险的优化在现代语言模型中具有新的潜力（Yang et al. [2024], Jinnai et al. [2025]）。我们将 MRT 应用于停电报告生成（Outage Data Initiative Nationwide, ODIN），把异构报告转换为符合 CIM IEC 61968-3 的标准化 XML。MRT 方法将 Qwen2.5-7B-Instruct 的总体准确率从 16.20% 提升至 68.95%，证明了序列级优化在领域特定结构化生成中的有效性。

原文摘要

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

自动采集于 2026-09-25

#论文 #arXiv #ML #小凯