Sichuan Conservatory of Music

Intangible Cultural Heritage (ICH) encompasses complex layers of symbolic meaning expressed through motifs, crafts, rituals, and regional traditions.Contemporary multimodal generative models frequently overlook such domain-specific semantics, leading to visually appealing but culturally inaccurate outputs.To address this limitation, this paper introduces a unified knowledge-graph-driven multimodal generation framework that couples a structured ICH Knowledge Graph (KG), a domain-adapted Large Language Model (LLM), and a controllable diffusion-based text-to-image generator.The KG organizes motifs, techniques, symbolic associations, and regional contexts into a structured semantic space, which the LLM leverages to interpret user queries and retrieve culturally grounded constraints.These constraints are injected into the diffusion model through a multi-stage semantic fusion mechanism, enabling culturally faithful and controllable image synthesis.Experimental results across three curated ICH datasets demonstrate that the proposed framework outperforms representative baselines in cultural semantic accuracy, text-image alignment, and robustness to linguistic variation.The proposed approach provides a principled pathway for integrating symbolic cultural knowledge with modern generative models, supporting large-scale preservation, computational interpretation, and creative revitalization of intangible cultural heritage.

ABSTRACTTraditional protection measures are difficult to cope with in complex and changeable network attacks. This paper presents an innovative construction of a hybrid model based on the combination of graph neural network (GNN) and random forest (RF), and uses an adaptive feature selection mechanism and multi‐task learning strategy to carry out network security big data prediction analysis. Through experiments on CICIDS 2017 and KDD Cup 1999 datasets, the results show that the model has excellent performance in detecting various types of attacks, with an accuracy of up to 0.94 in DoS attack detection of the CICIDS 2017 dataset, and maintains strong robustness in different noise environments. The model performs outstandingly in training efficiency, generalization ability, and multi‐task learning performance, and significantly improves the accuracy and efficiency of network security threat prediction. Research shows that this model has opened up a new path for network security protection, has important value in dealing with complex network attack scenarios, and provides strong support for the development of the network security field.

IntroductionIn China, Depression is a significant mental health issue among college students, with a 28.4% prevalence rate. This meta-analysis systematically synthesizes empirical data to evaluate the therapeutic efficacy of music therapy in mitigating depressive symptoms within collegiate populations.MethodsTwo authors searched seven international and Chinese databases from inception to February 5, 2025, for randomized controlled trials (RCTs) on music therapy for college students with depressive symptoms. Statistical synthesis was performed with RevMan 5.3 software, using random-effects models to calculate standardized mean differences (SMDs) with 95% confidence intervals.ResultsSixteen RCTs involving 938 participants were included. Pooled estimates showed significant reductions in depressive symptoms as measured by SDS (SMD = −1.19, 95% CI: −1.56 to −0.82; p < 0.001), SCL-90 depression subscale (SMD = −1.29, 95% CI: −1.96 to −0.62; p = 0.002), and BDI (SMD = −1.98, 95% CI: −3.59 to −0.38; p = 0.015). High heterogeneity (I2 = 76–89%) was found, but the results remained robust even after undergoing sensitivity analyses.ConclusionMusic therapy is an effective adjunctive intervention for managing depressive symptoms in college students. However, due to significant heterogeneity, future research needs standardized treatment frameworks. Multi-center RCTs with blinded assessment and longitudinal designs should be prioritized to clarify dose-response relationships and neurophysiological mechanisms, facilitating the integration of music therapy into evidence-based mental health strategies for this population.

As aging continues to deepen in Chengdu, community music activities, as an important vehicle for promoting the psychological health and social connection of older adults, and their influencing factors have become the focus of healthy aging research. Based on Maslow's hierarchy of needs theory and social-ecological system theory, this study conducted a questionnaire survey on a random sample of 480 older adults to systematically analyze the influence of individual needs, community environment, and social support on community music participation. The study found that: (1) demand-driven stratification: physiological needs and safety needs are the main driving forces for participation, while respect and self-actualization needs are seriously insufficient, leading to shallow participation behavior; (2) imbalance of environmental resources: the accessibility of community facilities, diversity of activities, and professional guidance are notably insufficient, especially in remote suburban communities; (3) fragile support system: family support is the core driving force, but the sustainability of activities is constrained by gaps in government funding and corporate sponsorship. The study proposes a synergistic path of “demand-environment-resource” optimization: activating high-level demand through layered activity design, breaking resource constraints through policy innovation and digital empowerment, and promoting the transformation of community music from “entertainment” to “value co-creation.” The theoretical value of this study is to build a localized framework for explaining cultural participation of the elderly and to provide a practical basis for decision-making on the construction of an “all-age friendly community” and the inheritance of NGT in Chengdu.

With the vigorous development of artificial intelligence (AI) in education, traditional piano teaching faces challenges such as low personalization, delayed feedback, and underutilization of classroom resources. Based on a systematic review of traditional piano teaching models and AI applications in education, and informed by cognitive psychology and intelligent teaching theories, this study employs a combination of questionnaires, in-depth interviews, and teaching experiments to analyze the current state and problems of piano classrooms. On this basis, we construct a restructured teaching model that integrates an intelligent teaching platform, personalized learning paths, and adaptive feedback, and conduct practical implementation and effectiveness evaluation in representative schools and student groups. The results show that this model significantly enhances student engagement, optimizes classroom management, and improves practice efficiency, while also revealing limitations in platform maturity and teacher training requirements. Finally, we discuss the feasibility and considerations for wider adoption of the model and propose future research directions in multimodal interaction and interdisciplinary integration.

Around the globe, technological advancements such as artificial intelligence (AI) are reshaping workplaces affecting employee wellbeing (EWB). To understand the AI-EWB link, a conceptual model is developed to explore the link between AI-driven capabilities and employee wellbeing (EWB), with cybernetic thinking (CT) as a mediator. Furthermore, organizational ambidexterity (OA) is introduced as a moderating factor between CT and EWB grounded on integrated dynamic capabilities with resource-based theory in the context of a developing country like Pakistan. Data were collected from 490 doctors working in private sector hospitals across two major cities of Pakistan—Karachi &amp; Islamabad and data analysis was performed using PLS-SEM 4.0. Results indicate that AI-driven capabilities significantly relate to EWB. Furthermore, CT explains the relationship between tangible, human resources, intangible-driven AI capabilities, and EWB. In addition, OA moderates the link between CT and EWB. Hence, mediated moderation is established. To remain resilient, this study offers theoretical as well as practical insights into how healthcare practitioners can harness AI through integrating organizational factors like CT can help reduce stress and improve EWB through adopting a balanced approach to manage innovation. Policy implications along with directions for studies to be conducted by researchers are also provided.

.

Instrument playing techniques (IPTs) constitute a pivotal component of musical expression. However, the development of automatic IPT detection methods suffers from limited labeled data and inherent class imbalance issues. In this paper, we propose to apply a self-supervised learning model pre-trained on large-scale unlabeled music data and finetune it on IPT detection tasks. This approach addresses data scarcity and class imbalance challenges. Recognizing the significance of pitch in capturing the nuances of IPTs and the importance of onset in locating IPT events, we investigate multi-task finetuning with pitch and onset detection as auxiliary tasks. Additionally, we apply a post-processing approach for event-level prediction, where an IPT activation initiates an event only if the onset output confirms an onset in that frame. Our method outperforms prior approaches in both frame-level and event-level metrics across multiple IPT benchmark datasets. Further experiments demonstrate the efficacy of multi-task finetuning on each IPT class. 1

The film, as a form of cultural expression, usually also assumes the responsibility of reflecting and exploring social issues and presenting them to the masses in a more popular way, and these social issues also include gender issues class issues, and so on. Feminism has been a hot topic in recent years, and it has led to many different issues such as women's independence and the dichotomy between men and women, but the public's understanding of feminism is too shallow and one-sided. Taking the film Thelma and Louise as an example, this study aims to explore the theme of women breaking the boundaries of gender and class to present feminist growth in this film, by analyzing the characters' personalities, their ways of thinking about things, and their final choices. This study likewise presents how the two female protagonists transform from victims to strong and independent women. According to the final results of the study Thelma and Louise successfully present a feminist coming-of-age story that is significant to society as well as the awakening of feminism.