no code implementations • CCL 2022 • Fengyuan Zhao, Dexi Liu, Qizhi Wan, Changxuan Wan, Xiping Liu, Guoqiong Liao
“现有的情感瘭原因对抽取模型均没有通过加入外部知识来提升情感瘭原因对的抽取效果。本文提出基于知识迁移的情感瘭原因对抽取模型瘨癅癃癐癅瘭癋癔瘩, 采用知识库获取文本的显性知识编码;随后引入外部情感分类语料库迁移得到子句的隐性知识编码;最后拼接两个知识编码, 加入情感瘨原因瘩子句预测概率及相对位置, 搭配癔癲癡癮癳癦癯癲癭癥癲机制融合上下文, 并采用窗口机制优化计算压力, 实现情感瘭原因对抽取。在癅癃癐癅数据集上的实验结果显示, 本文提出的方法超过当前最先进的模型癅癃癐癅瘭瘲癄。”
no code implementations • 21 Aug 2023 • Qi Chen, Dexi Liu
The combination of chain-of-thought (CoT) prompting and Large Language Models (LLMs) is employed and get the SOTA performance on various NLP tasks, especially on text generation tasks.
no code implementations • 4 Aug 2023 • Qi Chen, Dexi Liu
This innovative structure reduces the excessive reliance on pre-trained language models and emphasizes the modeling of structure and local relationships, thereby improving the performance of the model on Chinese financial texts.
no code implementations • 30 Jun 2023 • Qizhi Wan, Changxuan Wan, Keli Xiao, Hui Xiong, Dexi Liu, Xiping Liu
This paper introduces a novel framework for document-level event extraction, incorporating a new data structure called token-event-role and a multi-channel argument role prediction module.