Research And Application Of Chinese Text Matching Method Based On Information Interaction

Text matching is one of the research directions of natural language processing.Many natural language processing tasks can be abstracted as text matching problems,such as text deduplication,question answering system,etc.Traditional text matching methods mainly use matching operations based on lexical coincidence,which cannot understand the problem of text semantics.The text matching operation based on deep learning can effectively improve the accuracy of text matching,but it needs to process a large amount of data and requires a long computing time.The emergence of the pre-training model BERT has promoted great progress in natural language processing,and achieved remarkable results in multiple tasks,but it consumes a lot of resources and requires relatively high computing power.In the field of education,the distinction between professional terms and problems is not high,there are certain semantic diversity and text structure problems,and the requirements for text matching algorithms are relatively high.In order to solve the above problems,this paper adopts the text matching method based on information interaction,and proposes an information interaction-oriented Chinese text semantic matching model(IISM).Based on text retrieval,the model uses self-attention mechanism to extract deep semantic representations considering the context of the text itself.At the same time,the convolutional neural network is used to extract the structural information of the text,and in order to match the text to the low-level semantic information of the other party,the information of the word granularity and word granularity of the text is extracted,and the three kinds of semantic information are interacted to obtain a new semantic matrix.Bi LSTM performs feature collection and extraction.Then,the pre-trained model BERT is used in combination with the external language knowledge base to improve the word vector training of the model;the collaborative training algorithm is used to make the training performance of the model more stable;the particle swarm optimization algorithm is used to optimize the convolutional neural network hyperparameters.The model proposed in this paper has superiority in semantic information interaction and can solve the problems of semantic diversity and text structure in Chinese texts.In order to verify the effectiveness of the method proposed in this paper,experiments are carried out on the Chinese text matching dataset,and several commonly used text matching models are tested at the same time.The experimental results show that the Chinese text matching method based on information interaction proposed in this paper has a certain improvement in accuracy and precision,and has achieved good results,which has certain application value for the improvement and optimization of intelligent learning platforms.Applying this method to the intelligent answering system in the research and development project of artificial intelligence educational products in the laboratory can improve the performance of the intelligent answering system.