The application of machine learning in the field of music teaching is emerging, especially in piano teaching shows obvious potential. In this paper, we propose a key touch action recognition model based on ADAG-SVM. The Gaussian kernel function is chosen to solve the problem of high-dimensional vectors in the input space and feature space of action representation. And the optimal penalty parameters as well as the kernel radius are obtained to further improve the performance of touch-key action recognition. In addition, the piano timbre feature matrix is extracted, and based on the discrete Fourier transform, a tone synthesis model with editable timbre is established to generate expressive demonstration clips to deepen the students’ understanding of tonal expressiveness. The keystroke action recognition model in this paper can predict the change of the angle of the keystroke action of the student’s fingers, and provide scientific action guidance and correction for the learners. The LSD and MSD values of piano tones generated by this paper’s algorithm are 1.52 and 492.68 respectively, which are lower than those of the comparison algorithm. At the end of teaching, the scores of the C2 class under the intervention of this paper’s machine learning method in the piano level evaluation dimension improved by 0.62-1.97 points compared with the traditional teaching C1 class. Students’ satisfaction with the piano teaching class under this paper’s method ranged from 4.35 to 4.81 points, and the overall satisfaction reached 4.59.The piano teaching method combined with machine learning significantly improved students’ keystroke skills and tonal expression.
The wave of education informatization sweeps in, providing new opportunities and challenges for cultivating new talents. This study applies deep learning and learning theories related to combinatorial mathematics to college English teaching in order to improve the teaching effect, and builds a smart classroom teaching model of college English from three dimensions: guiding, advancing and strengthening. Using quantitative big data analysis methods, cluster analysis and correlation analysis of students’ learning behaviors are conducted to explore the group portrait of students in English courses. Then a teaching comparison experiment is conducted to verify and evaluate the application effect of the English smart classroom teaching model. The sample students were clustered into four types: excellent, diligent, maintenance and self-abandonment, with the diligent and maintenance types accounting for the highest proportion of students, 31.39% and 45.26%. The students’ English test scores and innovation cultivation effects as a whole improved by 5.65% and 16.24% respectively after the teaching experiment. The results show that the teaching design based on deep learning can effectively improve the teaching effect of college English smart classroom and promote the development of students’ innovative thinking and ability.
Integrating smart technology into the teaching of composition technology discipline can bring more innovative opportunities for the development of music education. In this paper, for the teaching of the theory course of composition technology, edge computing technology is combined with music education to propose a smart music teaching platform, and a knowledge tracking and exercise recommendation strategy is designed for this purpose. The strategy constructs a knowledge tracking model based on the self-attention mechanism and hypergraph module, on the basis of which the POMDP process is used to build exercise recommendation tasks to achieve personalized music exercise resource recommendation. Experiments show that the designed knowledge tracking model is more than 3% more accurate than the DKT method, and the mean value of each index of the exercise recommendation model is above 0.8, which is more targeted, novel and diversified. After the application of this teaching platform, students’ attitudes and willingness towards the course, interaction and platform have increased significantly, by 25.25%, 30.53% and 37.68% respectively. Meanwhile, more than 60% of the students affirmed its enhancing effect on teacher-student interaction, exercise recommendation, course interest and learning effect. The proposed teaching platform effectively promotes teacher-student interaction teaching, provides personalized aids for music course learning, and can help promote the development of music teaching quality in colleges and universities.
The application of digitalization and intelligence-related technologies brings unprecedented opportunities and challenges to the development of university English translation teaching. In response to the current English teaching reform, this paper designs a university English translation teaching model based on mathematical multiple intelligence theory. In order to make this teaching mode better serve students and teachers, a personalized recommendation algorithm based on Markov chain is introduced on the original teaching mode. The research samples are selected, the groups as well as the questionnaires are set, the research hypotheses and methods are given, and the task of formulating the research program is finally completed. With the help of independent samples t-test, the university English translation teaching model of this paper is validated and analyzed. There are significant differences between the two in comprehension (F=2.421, p=0.009<0.01), analytical (F=16.522, p=0.002<0.01), practical (F=6.928, p=0.008<0.05), and reflective (F=8.006, p=0.009<0.01) skills, which indicates that the model of university English translation teaching based on the Markov chain and the Mathematical Multiple Intelligences Theory of University English Translation Teaching Model, which has a more obvious effect on the improvement of students’ English translation ability.
In recent years, music has been developing rapidly in China and has become one of the hot items of current school teaching, and more and more schools have carried out college music education. However, there are still many problems that need to be solved in college music education, among which the problem of unreasonable teaching methods is particularly prominent. In this study, we use federal reinforcement learning technology to explore in depth the theory and method of the situational classroom teaching model and implement it into music education in China’s general colleges. Through empirical testing, we further analyze the optimization effect and empirical variability of the model to provide reference for the implementation and promotion of the situational classroom teaching model in China’s general college music education. The experimental results proved that the optimized federal average algorithm achieved better results than the traditional teaching method, using the situational classroom teaching method in college music education, which could improve the quality of set completion and significantly improve the teaching quality, and the accuracy of the optimized model of federal learning increased by 8%.
The traditional English teaching mode ignores the differences between individual students, resulting in some students’ low motivation to learn, while personalized English learning based on the support of Internet of Things can effectively improve students’ learning efficiency. In this paper, we construct the evaluation index system of student engagement from four dimensions: learning attitude, learning process, learning effectiveness and emotional experience, and quantify the degree of student engagement in IoT-supported personalized English learning through a fuzzy mathematical model. In order to further improve student engagement, a student engagement improvement mechanism is designed based on four evaluation results: excellent, good, moderate and poor. The study takes the freshman (1) class of English majors in School A, which uses IoT technology for personalized English learning, as the research object and uses a fuzzy mathematical model to quantify student engagement. The evaluation grade of student engagement in this class is “medium”, and the evaluation score is 71.37. Accordingly, a student engagement improvement mechanism was introduced. After 8 weeks of rectification, the final grade of student engagement in English learning is “excellent”, with an evaluation score of 88.49, which is a total improvement of 17.12 points, and the student engagement enhancement mechanism designed in this paper has a significant effect.
In the context of contemporary higher education, campus space design serves as a vital tool for teacher student interaction and the development of a sense of community in addition to serving the needs of studying and living. However, social contact and psychological belonging are frequently overlooked in traditional campus design, creating a gap between physical arrangement and real demands. This study offers a number of creative approaches for managing and designing campus spaces that are based on the student community interaction paradigm. Teaching, housing, and community services are organically integrated to improve the interaction between teachers and students through the introduction of the “molecular unit” concept. Resources are shared and more opportunities for social practice are realized through the establishment of a comprehensive mechanism between the campus and the urban society. At the same time, the psychological identity and sense of belonging of the teachers and students are enhanced through the construction of the event space and the spirit of the place. Furthermore, the idea of hybrid living that this study promotes can aid in removing barriers between students and instructors and further improve campus life. The results encourage spatial innovation and cultural growth in higher education settings and offer theoretical support and useful guidance for future campus design.
The rapid economic development, coupled with accelerated industrialization and urbanization, has brought significant challenges to sustainable development due to environmental pollution and climate change. Promoting the adoption of eco-friendly home products is essential for mitigating household pollution and advancing green, low-carbon, and sustainable consumption. This study investigates the key factors influencing consumers’ purchasing behavior of eco-friendly home products and proposes targeted strategies. Grounded in the Theory of Planned Behavior (TPB) and the “Attitude- Intention-Behavior” model, the research constructs a theoretical framework for understanding green consumption behavioral intentions. Data were collected through an online survey of 397 consumers of home products in the Sichuan-Chongqing region, and Structural Equation Modeling (SEM) was employed to empirically examine the determinants of consumers’ eco-friendly purchasing behavior. The results demonstrate that the proposed model effectively explains and predicts consumers’ purchasing decisions regarding eco-friendly home products. Among the identified factors, attitude toward green products is the most significant determinant, while environmental concern and environmental awareness are the strongest antecedents influencing purchase intentions. Perceived behavioral control also exhibits a positive and significant effect on purchase intentions, whereas perceived environmental knowledge shows no significant impact. These findings offer valuable theoretical insights for manufacturers and market practitioners aiming to better align with and influence consumer decision-making in the ecofriendly home products sector. Additionally, the study provides practical recommendations for policymakers focused on fostering green consumption and promoting the sustainable development of eco-friendly home products.
Spectral variability remains a major challenge in hyperspectral unmixing, as the spectral signatures of the same material often fluctuate under different illumination, atmospheric, or scale conditions, which invalidates the fixed endmember assumption. To address this issue, we propose Meta-Spectral Net, a meta-learning-based framework for adaptive spectral variability modeling in hyperspectral unmixing. The proposed framework leverages a task-driven meta-learning strategy, where each acquisition scenario is defined as a task, to enable the endmember generator to rapidly adapt to unseen spectral conditions with only a few samples. Furthermore, a spectral variability adaptation module is introduced to explicitly account for environmental factors, thus improving the robustness of endmember representation. Comprehensive experiments on both synthetic and real hyperspectral datasets demonstrate that Meta-Spectral Net significantly outperforms state-of-the-art unmixing methods in terms of endmember reconstruction accuracy and abundance estimation, while offering superior generalization to novel scenarios. These results suggest that meta-learning provides an effective paradigm for tackling spectral variability, paving the way toward more adaptive and reliable hyperspectral unmixing in real-world applications.
The rapid development of the digital economy and the deployment of large-scale electricity big-data platforms have highlighted both the opportunities and risks associated with energy data circulation. Conventional identity management frameworks in the energy sector suffer from weak authentication, fragmented governance, and high compliance costs, limiting the secure and efficient realization of data value. This paper proposes a decentralized digital identity (DID) management framework tailored for the energy data space. By integrating blockchain-based traceability, verifiable credentials, and smart-contract-driven privacy protection, the framework establishes a sovereign, interoperable, and privacy-preserving identity infrastructure. Through simulation experiments using Monte Carlo modeling, we evaluate the performance of the proposed system under different blockchain infrastructures (Fabric vs. EVM) and disclosure mechanisms (plain vs. zero-knowledge proofs). The results demonstrate that Fabric achieves lower latency and higher throughput compared to EVM, while zero-knowledge proofs introduce moderate but acceptable overhead, enabling stronger privacy guarantees. The proposed framework effectively tackles the key challenges of secure identity verification, fine-grained and dynamic authorization, and tamper-resistant auditing, thereby establishing a scalable and reliable foundation for trusted circulation of energy data.