用课程学习优化大模型，生成更可信的印地语新闻事实解释
From Fragments to Facts: A Curriculum-Driven DPO Approach for Generating Hindi News Veracity Explanations
- 结合课程学习与直接偏好优化，让模型逐步学会像人一样推理
- 引入真实度和细腻度两个参数，使解释更准确一致
- 特别适合低资源语言如印地语的假新闻分析，可扩展至其他语言
在虚假信息泛滥的时代，生成可靠新闻解释至关重要，尤其对印地语等代表性不足的语言。由于缺乏强大的自动化工具，印地语在假信息检测方面面临挑战。为此，我们提出一种新框架，将直接偏好优化（DPO）与课程学习相结合，使机器生成的解释与人类推理对齐。来自权威信源的事实核查解释作为优选响应，而大模型输出则反映系统局限性，作为非优选响应。为提升任务特定对齐，我们在DPO损失函数中引入两个关键参数——真实度（Actuality）与细腻度（Finesse），以增强解释的质量与一致性。实验使用Mistral、Llama、Gemma等大语言模型及mBART、mT5等预训练语言模型，验证了该框架在生成连贯、上下文相关解释方面的有效性。该可扩展方法有助于对抗虚假信息，并将自动化解释生成推广至低资源语言。
原文摘要 · Abstract (English)
In an era of rampant misinformation, generating reliable news explanations is vital, especially for under-represented languages like Hindi. Lacking robust automated tools, Hindi faces challenges in scaling misinformation detection. To bridge this gap, we propose a novel framework integrating Direct Preference Optimization (DPO) with curriculum learning to align machine-generated explanations with human reasoning. Fact-checked explanations from credible sources serve as preferred responses, while LLM outputs highlight system limitations and serve as non-preferred responses. To refine task-specific alignment, we introduce two key parameters -- Actuality and Finesse -- into the DPO loss function, enhancing explanation quality and consistency. Experiments with LLMs (Mistral, Llama, Gemma) and PLMs (mBART, mT5) confirm the framework's effectiveness in generating coherent, contextually relevant explanations. This scalable approach combats misinformation and extends automated explanation generation to low-resource languages.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。