CN 116992979 A

Abstract

本发明公开了一种基于深度学习的认知训练任务推送方法及系统。该方法包括如下步骤：获取用户画像数据和认知训练任务数据，并输入预训练模型；基于预训练模型进行预处理，形成预处理数据；基于预训练模型对预处理数据进行特征嵌入和特征拼接，形成输入向量；基于预训练模型对输入向量进行加权求和与深度兴趣网络分析，输出加权求和结果和深层次特征信息；基于加权求和结果，预测认知训练任务得分；并基于深层次特征信息，预测认知训练任务组合的任务得分增量列表；重复上述过程，以不断预测用户不同认知训练任务组合的任务得分增量列表，从而选取增量变化最大的认知训练任务组合推送给用户进行认知训练。

English (machine translation): The invention discloses a deep learning-based cognitive training task pushing method and system. The method comprises the following steps: acquiring user portrait data and cognitive training task data, and inputting a pre-training model; preprocessing based on the pre-training model to form preprocessed data; performing feature embedding and feature splicing on the preprocessed data based on the pre-training model to form an input vector; carrying out weighted summation and deep interest network analysis on the input vector based on the pre-training model, and outputting a weighted summation r…

The opening of the abstract. The full text is in the PDF.