| Nowadays,artificial intelligence,serves as significant part in our society,raises increasing attractiveness and reliability from human.Advancing for efficiency increasing and information searching accuracy improving,as the part of artificial intelligence technology,the text correction tool is widely employed in various applications in our daily.However,the complex Chinese syntax and over-parameterization of the language model greatly challenges the Chinese text correction method facing the actual application requirements.Specifically,the obstacles including:1.Aware of the processing level of the correction method,previous works are divided as Chinese spelling correction and Chinese grammatical error correction.Within multiple meanings of the words under different contexts,the correction methods with various processing level will lead to inconsistent correction results,which results in unreliable for correction process.2.While explosive parameter-increasing of the language models improves the performance of text correction,the over-parameterization blocks the model decision process from interpretability.The black-box correction process will lead to ethical issues for machine,restrict the technology application from important scenarios and raise unreliability between human and machine.In order to tackle the above obstacles,we proposed two difference methods in this paper:1.An adversarial multi-task learning method is proposed to enhance the modeling and detection ability of character polysemy in Chinese sentence context.Wherein,two models,the masked language model and scoring language model,are introduced as a pair of not only coupled but also adversarial learning tasks.Moreover,the Monte Carlo tree search strategy and a policy network are introduced to accomplish the efficient Chinese text correction task with semantic detection.The experiments are executed on three datasets and five comparable methods,and the experimental results show that our method can obtain good performance in Chinese text correction task for better semantic rationality.2.A novel interpretable deep learning model(AxBERT)is proposed for Chinese spelling correction by aligning with an associative knowledge network(AKN).Wherein AKN is constructed based on the co-occurrence relations among Chinese characters,which denotes the interpretable statistic logic contrasted with uninterpretable BERT logic.And a translator matrix between BERT and AKN is introduced for the alignment and regulation of the attention component in BERT.In addition,a weight regulator is designed to adjust the attention distributions in BERT to appropriately model the sentence semantics.Experimental results on SIGHAN datasets demonstrate that AxBERT can achieve extraordinary performance,especially upon model precision compared to baselines.Our interpretable analysis,together with qualitative reasoning,can effectively illustrate the interpretability of AxBERT.3.After revisiting the coupled requirement of deep neural representation and semantics logic of language modeling,a Character-Context-Coupled Space(C3Space)is proposed by introducing the alignment processing between uninterpretable neural representation and interpretable statistical logic.Moreover,a clustering process is also designed to connect the character-and context-level semantics.Specifically,an associative knowledge network(AKN),considered interpretable statistical logic,is introduced in the alignment process for character-level semantics.Furthermore,the context-relative distance is employed as the semantic feature for the downstream classifier,which is greatly different from the current uninterpretable semantic representations of pre-trained models.Our experiments for performance evaluation and interpretable analysis are executed on several types of datasets,including SIGHAN,Weibo,and Chn Senti.Wherein a novel evaluation strategy for the interpretability of machine learning models is first proposed.According to the experimental results,our language model can achieve extraordinary performance and highly credible interpretable ability compared to related state-of-the-art methods.4.Additionally,this paper conducted tests and analyses on the performance of ChatGPT series models,which have had a huge impact on natural language processing field,in spelling correction and sentiment analysis tasks.The results show that these models have powerful logical reasoning and semantic understanding capabilities.However,when it comes to executing the given task requirements,their execution and completion are relatively weak. |
