Acupuncture has been recognized by more and more experts as a treatment method to relieve various pains in human body, but the association between specific acupuncture treatments and diseases is still unclear, which affects the long-term development of acupuncture treatment. In this paper, we abstract the knowledge of acupuncture points as ontologies in the knowledge graph, and propose a method to improve the RoBERTa-WWM-BiGRU-CRF model to optimize the knowledge extraction of the knowledge graph by combining the SoftLexicon technique and the adversarial training method. Based on the knowledge graph of acupuncture points, the collaborative filtering model is introduced, and the original similarity matrix construction method is replaced by the co-occurrence matrix construction method based on the association characteristics of acupuncture points and diseases, which improves the operational efficiency of the association search and realizes the design of the association search technology of acupuncture points and diseases. The average consultation time in the acupuncture outpatient departments of the experimental and control groups applying this paper’s technology for acupuncture visits was faster than that of the full outpatient clinic by 0.32 min, showing a significant difference (P<0.05). Patients in the experimental group who received acupuncture treatment assisted by the technology of this paper were higher than those in the control group in the dimensions of acupuncture treatment experience, such as physiological reflections, treatment emotions, and treatment effects and treatment feeling dimensions, which were 2.22, 3.57, 2.2, and 1.33, respectively.

In-depth investigation of the combination of innovation and entrepreneurship education and computing technology is of great theoretical and practical significance for the continuous promotion of innovation and entrepreneurship education in colleges and universities. In this paper, after clarifying the three elements of environment, subject and behavior in the design of innovation and entrepreneurship education courses, we design an innovative teaching model for innovation and entrepreneurship education courses based on computing technology and digital learning environment, and adopt similarity metrics and questionnaires to count the frequency of students’ on-line learning behaviors and the level of cultivation of their innovative spirit and ability respectively. The results of teaching practice show that the practice of innovation and entrepreneurship education courses based on computing has a facilitating effect on the cultivation of students’ innovative spirit and ability. The Spearman correlation coefficients of the learning behaviors in the online teaching platform of innovation and entrepreneurship education courses and the dimensions of innovation spirit and ability show medium-high correlation (r>0.3), and its regression model can effectively explain more than 60% of the variance of innovation spirit and ability. The research in this paper provides an effective reference for the innovative development and practice of innovation and entrepreneurship education programs, and lays the foundation for promoting more effective and innovative development of dual innovation education in colleges and universities.

Natural language processing (NLP) is developing very rapidly in the field of artificial intelligence, and has become an important direction in the development of computer science field and artificial intelligence industry. In this paper, in order to realize the efficient construction of natural language processing model in low-dimensional embedding space, firstly, a word vector learning model is constructed based on matrix decomposition for word vectors in natural language processing. On this basis, in order to further realize the efficient construction of natural language processing models, this paper designs the Semantic Discarding Network (SDN) and Semantic Fusion Alignment Method (SFA) for the problem of interfering semantics of the model and the problem of a single way of fusion of local inference results. Finally, the SDF-NN natural language processing model is proposed and a multi-view subspace clustering (DLTE) method based on deep low-rank tensor embedding is proposed. The results of the research experiments show that the average performance index of this paper’s word vector model for each task in three corpora ranges from 71.55 to 89.11, and the performance is stable and the time overhead in the three corpora is 3.93, 7.29, and 13.42 minutes, respectively, and the speed of the model has been significantly improved and the overall performance is better. In addition, the natural language processing model (SDF-NN) constructed in this paper achieves the best performance in the comparison test with strong competitiveness, which further validates the performance of the matrix decomposition-based natural language processing model in this paper, and provides the method and direction for its efficient construction in low-dimensional embedding space.

The risk of financial aspects intuitively reflects the development status and operating results of enterprises, enterprises must control the financial risk of this key link, so that the financial risk of a safe landing, to protect the stability and health of the enterprise. This paper selects the financial data of listed companies, and comprehensively analyzes the level of the company’s financial performance from four aspects, namely, profitability, operating capacity, growth capacity and solvency indicators. Using Benford’s law to test the quality of each data of each financial indicator, the Benford factor is introduced as a new explanatory variable, and combined with the company’s financial risk early warning indicators to establish a random forest early warning model. The results show that profitability and growth capacity are the strengths of listed companies, while operational capacity and solvency are the weaknesses. The results analyzed by K-means clustering algorithm show that the sample companies are divided into 5 categories. And compared with the basic random forest model, the random forest model based on Benford’s law can improve the accuracy of financial risk warning. Finally, the model with the best prediction effect is used to judge the financial status of G listed companies, get the early warning results, verify the accuracy and applicability of the model and put forward corresponding countermeasure suggestions.

The development and utilization of shale gas is the main path to solve the current high carbon dioxide emissions, and this paper proposes to use the LEAP model to explore the role of shale gas development and utilization on carbon emission reduction in all aspects. Under the principle and definition of LEAP model framework, shale gas development and utilization scenarios and parameters are determined to facilitate the research and analysis work, and in order to realize the intelligent monitoring of carbon emission reduction work, the neural network two-layer carbon emission reduction prediction model is constructed. With the support of research data and LEAP model, the relationship between shale gas development and utilization and carbon emission reduction is studied and analyzed, and the carbon pulse analysis and prediction model validation model of LEAR simulation results are also supplemented. Although all three scenarios have different contributions to carbon emission reduction, the green scenario is the most obvious means of carbon dioxide emission reduction, with a total of 52.87 from 2010 to 2050, and the prediction model in this paper is able to satisfy the current demand for carbon dioxide emission reduction work, and provide a guiding reference for urban carbon emission reduction.

The rapid development of artificial intelligence technology has made its application in the field of education increasingly widespread. The purpose of this paper is to design and implement a personalized vocal music teaching system based on artificial intelligence algorithms to solve the problems of single teaching method and lack of personalized guidance that exist in traditional vocal music teaching. The overall architecture of the system is constructed by analyzing the demand for vocal music teaching and combining deep learning and other artificial intelligence technologies. The key algorithms involved in the system are elaborated in detail, including the personalized recommendation algorithm of the learning path fused with the long and short-term memory network (LSTM) and the attention mechanism, and the intelligent evaluation algorithm that includes the evaluation of pitch, rhythm and timbre. Through practical application cases, it is verified that the system in this paper can effectively improve the teaching effect of vocal music and students’ vocal music professionalism, providing an important auxiliary role and key ideas for the innovative development of vocal music teaching.

This study firstly introduces the working principle of deep learning-based neural machine translation model (NMT) and its recurrent neural network translation backbone network, which enhances the semantic characterization capability through Glove word embedding layer. A tree-to-sequence based attention mechanism is innovatively introduced at the encoder side, and a tree-based encoder is appended to the traditional sequence encoder to construct syntax-aware context vectors. On the decoder side, the syntactic tree structure information is integrated into the sequence-to-sequence model (seq2seq), and this model is used to explore the knowledge transfer effect of the English translation teaching process. The results show that the accuracy rates of the neural machine English translation models incorporating syntactic information proposed in this paper are all above 90%. The experiment on the effect of English translation teaching shows that the mean values of students’ scores on the post-test of long sentence translation and composition translation in the reading section of the experimental class increased by 11.022 and 12.5388 points respectively compared with those of the control class, with significant differences between the scores of the two groups of students (p<0.05), and the same significant differences are presented between the scores on the pre-test and post-test of the students' scores on the long sentence translation and composition translation in the experimental class. It can be seen that the application of the model can effectively promote knowledge transfer and help students better understand and utilize translation skills.

In order to realize the intelligent calculation of cost management during the implementation of construction projects, this paper proposes a methodological architecture based on Multi-intelligent Reinforcement Learning (MARL) and Building Information Model BIM. The construction cost management problem of the project is analyzed with examples in order to optimize the construction cost management and construction time management of ZZYH comprehensive business building. The results of the study show that a reasonable rebar path can be found through 40 independent simulation verifications, and the final convergence reaches 100%. Compared to manual savings, the collision-free rebar design using the computational framework of BIM and multi-intelligence saves roughly 90% of the time. In terms of optimizing the construction cost management of civil engineering, installation engineering, cable engineering, and overhead line engineering, the total amount of cost savings of the project amounted to 382,320,000 yuan.

Based on the material demand forecasting model using BP neural network and particle swarm algorithm, the study builds the material whole chain response efficiency calculation model under dynamic multi-objective optimization by comprehensively considering the demand level weights of the affected area, and adopts genetic algorithm to assist the model solution in finding the optimal and decision-making. Taking an earthquake as a case for example analysis, the model in this paper can give the Pareto frontier, and combined with the weight coefficients after the transformation of the model solving results are more scientific and feasible, the demand satisfaction rate of the original model and the transformed model are 73.43% and 74.28% respectively, and the demand satisfaction rate of the affected points is improved by 4.24%, and this paper introduces the material allocation model of the demand level weights to be able to obtain better response efficiency of the whole chain of materials, which can provide important theoretical and practical guidance for the whole chain distribution of materials.

With the rapid development of digital technology, the inheritance and dissemination of folklore sports culture have ushered in new opportunities and challenges. This paper constructs a digital educational resource management platform for China-ASEAN folklore sports culture based on Knowledge Graph. The knowledge system of folklore sports culture is systematically constructed by using Knowledge Graph, the data related to China-ASEAN folklore sports culture are collected and organized, and the construction of the corpus of China-ASEAN folklore sports culture domain is completed. Then we extracted knowledge from the data of folklore sports culture domain and stored the obtained knowledge in Neo4j graph database. The China-ASEAN Folklore Sports Culture Digital Education Resource Management Platform, which includes several modules such as login and registration, courses, personal center, institutions and teachers, and backstage management, was designed. The construction and application of the platform gained 91.2% satisfaction from students, enhanced students’ interest in learning folklore sports culture, helped to protect and pass on the rich China-ASEAN folklore sports cultural heritage, and also promoted in-depth exchanges and communication between the two sides in the field of sports and culture to build a community of human destiny.