In recent years, research on the automatic input of graphics became active, because researches in the fields of CAD, GIS (geographical information system), OA (office automation), etc. have raised high requirements fo...In recent years, research on the automatic input of graphics became active, because researches in the fields of CAD, GIS (geographical information system), OA (office automation), etc. have raised high requirements for the graphic input of computers. But compared with the graphic output approach, the current graphic input approach is so limited that it cannot meet the needs of intelligent input.展开更多
针对电网调度业务意图缺乏有效识别方法的问题,提出一种基于ALBERT(A Lite BERT)和残差向量-字词嵌入向量-编码向量(RE2)融合模型的电网调度意图识别方法。首先,基于ALBERT预训练的动态词向量计算调度专业语言文本特征,建立调度意图分...针对电网调度业务意图缺乏有效识别方法的问题,提出一种基于ALBERT(A Lite BERT)和残差向量-字词嵌入向量-编码向量(RE2)融合模型的电网调度意图识别方法。首先,基于ALBERT预训练的动态词向量计算调度专业语言文本特征,建立调度意图分类模型,通过训练调度专业语言构建基于RE2的文本相似度计算模型。然后,采用RE2相似度模型计算召回文本与分类文本的匹配结果对ALBERT意图分类权重进行计算重组,建立融合ALBERT和RE2的意图识别模型。最后,通过某调控中心调度专业语言验证,并与其他方法对比,所提电网调度意图识别方法具有更强的分类能力和泛化能力,对于20种调度意图识别的平均精准率、召回率和F1值分别达到了98.11%、97.96%、98.03%。展开更多
基金Project supported in part by the 863 Plan and in part by the National Natural Science Foundation of China
文摘In recent years, research on the automatic input of graphics became active, because researches in the fields of CAD, GIS (geographical information system), OA (office automation), etc. have raised high requirements for the graphic input of computers. But compared with the graphic output approach, the current graphic input approach is so limited that it cannot meet the needs of intelligent input.
文摘针对电网调度业务意图缺乏有效识别方法的问题,提出一种基于ALBERT(A Lite BERT)和残差向量-字词嵌入向量-编码向量(RE2)融合模型的电网调度意图识别方法。首先,基于ALBERT预训练的动态词向量计算调度专业语言文本特征,建立调度意图分类模型,通过训练调度专业语言构建基于RE2的文本相似度计算模型。然后,采用RE2相似度模型计算召回文本与分类文本的匹配结果对ALBERT意图分类权重进行计算重组,建立融合ALBERT和RE2的意图识别模型。最后,通过某调控中心调度专业语言验证,并与其他方法对比,所提电网调度意图识别方法具有更强的分类能力和泛化能力,对于20种调度意图识别的平均精准率、召回率和F1值分别达到了98.11%、97.96%、98.03%。
