发表机构
Arizona State University（亚利桑那州立大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本文提出EqGrid模拟系统，采用基于现实场景的LLM策略智能体，可降低能源贫困公平性指标的不平等，压缩后的小参数模型仍能保留大部分收益，且计算效率大幅提升。
AI 中文摘要
能源贫困问题在面向社会公益的自然语言处理（NLP）研究中几乎未被关注，现有少量相关工作要么是静态检索/问答任务，要么依赖高碳排放的云端大语言模型（LLM），这种情况在人道主义场景中是一种适得其反的“计算讽刺”。本文提出EqGrid，这是一个闭环模拟系统，其中低频、开放权重的LLM策略智能体针对具有现实基础的家庭角色设定价格、碳排放上限及定向补贴，而高频多智能体强化学习（RL）交易者则在受物理配电网约束的连续双重拍卖中完成清算（该电网为带有动态运行 envelopes 的IEEE 33节点母线系统）。本文的贡献有三方面，且直接解决如何衡量AI的社会影响：（i）基于现实场景的角色（匹配区域的社会人口统计学特征），其负荷曲线的形状和水平已对照真实智能电表数据验证了真实性；（ii）正式的能源贫困公平性指标（能源负担、能源负担的基尼系数、低收入家庭社区（LIHC）指标），结果显示该干预措施可降低负担不平等，且不会增加电网总成本；（iii）计算效率前沿，用于衡量将策略智能体从2350亿参数的教师模型压缩到可在笔记本电脑上部署的10亿参数以下模型时，公平性能的保留程度，以及每次决策的估计能耗/碳排放量。解耦安全设计（LLM设定上限，验证与投影电网门执行）实现了零电网约束违反，而直接由LLM控制时违反次数为55次。在能源贫困公平性方面，LLM策略将能源负担基尼系数从0.351降至0.305，平均负担降低28%，同时降低了成本（优于调优后的规则基线）；30亿激活参数模型保留了95%的收益，推理能耗约为教师模型的9倍；即使是0.8亿参数的设备端模型也保留了92%的收益，能耗约为教师模型的24倍。本文将发布代码和配置文件。
英文摘要
Energy poverty is nearly absent from NLP-for-social-good, and the little existing work is either static retrieval/QA or relies on carbon-intensive cloud LLMs, a self-defeating "computational irony" for a humanitarian setting. We present EqGrid, a closed-loop simulation in which a low-frequency, open-weight LLM policy agent sets price and carbon bounds and targeted subsidies over a community of empirically-grounded household personas, while high-frequency multi-agent RL traders clear a continuous double auction constrained by a physical distribution grid (IEEE-33-bus with Dynamic Operating Envelopes). Our contribution is threefold and directly addresses how to measure the social impact of AI: (i) grounded personas (region-matched socio-demographics) whose load curves are checked for shape and level realism against real smart-meter data; (ii) formal energy-poverty equity metrics (Energy Burden, Gini of EB, LIHC) showing the intervention reduces burden inequality without raising net grid cost; and (iii) a compute-efficiency frontier that measures how much equity performance survives compressing the policy agent from a 235B teacher down to a sub-1B model deployable on a laptop, in estimated energy/carbon per decision. A decoupled-safety design (the LLM sets bounds; a validate-and-project grid gate executes) yields zero grid-constraint violations versus 55 under direct LLM control. On energy-poverty equity, the LLM policy lowers the Gini of energy burden to 0.305 (from 0.351) and mean burden by 28% while cutting cost (outperforming a tuned rule baseline), and a 3B-active model retains 95% of the benefit at roughly 9x lower inference energy than the teacher, with even a 0.8B on-device model retaining 92% at roughly 24x lower energy. We will release code and configs.
Comments9 pages, 2 figures, 4 tables. Accepted to the 5th Workshop on NLP for Positive Impact (NLP4PI) at EMNLP 2026
