面向多标签文本分类的标签依赖感知集预测网络（英文版）

英文标题：Label Dependencies-aware Set Prediction Networks for Multi-label Text Classification中文摘要：通过构建邻接矩阵及应用 GCN 模型对标签进行建模，从而解决多标签文本分类中标签的序列乱序问题，并利用集合预测网络同时使用句子信息和标签信息进行分类。此外，使用巴氏距离对输出概率分布进行约束，提高了召回能力。实验证明，该方法在多个数据集上表现优异，胜过之前的方法。英文摘要：Multi-label text classification aims to extract all the related labels from asentence, which can be viewed as a sequence generation problem. However, thelabels in training dataset are unordered. We propose to treat it as a directset prediction problem and don't need t