Word embedding method and device, and electronic equipment

The invention provides a word embedding method, which comprises the steps of receiving word vector data through a preset neural network, the preset neural network comprising a generator and a decisiondevice; extracting a feature tensor from the word vector data through the generator, wherein the generator is formed based on a deconvolution network; and classifying the feature tensors through thedecision device to output a classification result, wherein the decision device is formed based on a convolutional network. According to the word embedding method, the number of adjustable parameters of the model can be effectively reduced, the depth of the model is increased, and the generalization ability of the model is improved. 本公开提供一种词嵌入方法，包括：通过预设神经网络接收词向量数据，所述预设神经网络包括生成器和判决器；通过所述生成器对所述词向量数据提取特征张量，所述生成器基于反卷积网络形成；通过所述判决器对所述特征张量进行分类以输出分类结果，所述判决器基于卷积网络形成。本公开提供的词嵌入方法可以有效减少模型的可调整参数数量，增加模型深度，提高模型泛化能力。