Beyond traditional user recommendation, group recommendation is a new and popular task that provides recommendations for a group of users. Previous works aggregate member preferences in the group to infer group preference, but this often leads to a coarse-grained inference for group preferences limited by users’ individual preferences. To this end, we exploit that user preferences can be inferred and refined by exploring the group preferences that they participated in. These refined preferences offer additional information beyond the original individual preferences, enabling more fine-grained and satisfactory group preference inference. In this work, we propose a novel D ual- G ranularity G raph V ariational A uto- E ncoder framework (DGGVAE) for group recommendation, which jointly reveals group preferences from both coarse granularity and fine granularity to comprehensively learn group preferences. Specifically, we design a Group Preference Extractor module that extracts group preferences from these two granularities: coarse granularity, which is revealed through original member preferences, and fine granularity, which is revealed through refined member preferences. To extract the correlation between groups, a Group Representation Enhancement module is proposed, which enhances group representations by information from the most similar groups. However, the coarse- and fine-grained group preferences contain uncertainty due to the gap between the original and refined member preferences. To better incorporate dual-granularity group preferences, we design granularity-specific graph variational encoders that learn Gaussian variables on the semantic information for each group. Moreover, with the conditional independence assumption, the granularity-specific Gaussian node embeddings are fused according to the generalized product-of-experts (gPoE), where the semantic information in each granularity is weighted based on the estimated uncertainty level. Extensive experiments show the superiority of DGGVAE over various state-of-the-art methods in training efficiency and accuracy on both group and user recommendation tasks.
The increasing consumption of plastic in the world has led to severe environmental problems, particularly in the disposal of polyethylene terephthalate (PET) waste. In this regard, this paper discusses the feasibility of incorporating waste PET fibers in high-strength reinforced concrete (HSC RC) beams to enhance shear performance and promote the utilization of sustainable construction materials. Because HSC is inherently brittle, it can suddenly fail due to shear. Therefore, strategies to increase its ductility and energy absorption capacity are required. Despite reports of PET fibers improving tensile and flexural properties, the effect of PET fibers on shear strength is not clearly known. While some studies have explored PET fibers in normal-strength concrete, there is no data on the optimized interaction between specific fiber lengths and high-volume fractions in High-Strength Concrete (HSC) beams failing in shear. This study complements existing literature by evaluating the effect of PET fiber length (30 mm, 40 mm) and volume fractions (up to 1.5
The digital preservation of intangible cultural heritage (ICH) has become a vital strategy for sustaining cultural diversity in the face of globalization and digital transformation. This study employs bibliometric analysis and CiteSpace visualization tools to examine 798 research articles from the Web of Science (WoS) and China National Knowledge Infrastructure (CNKI) databases (2006–2024). It provides a systematic comparative analysis of methodologies and paradigms in ICH digitization across Chinese and Western academic discourses. The findings reveal distinct conceptual orientations: Western research tends to be technology-centric, emphasizing virtual reality (VR), augmented reality (AR), and blockchain for heritage data modeling, digital archiving, and virtual exhibitions. In contrast, Chinese research adopts a culture–technology symbiosis approach, focusing on digital storytelling, with an emphasis on tourism-integrated innovation, community participation, and interdisciplinary collaboration. A temporal analysis highlights digital twins and artificial intelligence (AI) as emerging transformative forces shaping global ICH preservation, integrating technological advancements with cultural imperatives. To address existing gaps, this study proposes a technology–culture–community synergy framework, fostering a holistic and inclusive approach to ICH digitization. By bridging technological rationality with humanistic values, this study contributes to comparative ICH digitization discourse and provides practical guidance for building inclusive digital ecosystems. Ultimately, it contributes to global heritage discourse by redefining digital technology not only as a preservation tool but also as a driver of cultural innovation, offering a strategic roadmap for balancing technological advancement with the ontological integrity of intangible heritage.
The role of GenAI in learners' emotional engagement has become an increasingly relevant topic in recent times, particularly in how it shapes learners' interactions with and processing of written feedback. The present study explores the effectiveness of peer feedback compared to GenAI feedback in EFL writing context. Participants (n = 152) were first-year undergraduate students at a Chinese university enrolled in an English writing course. The peer feedback group (n = 78) received feedback from peers, while the GenAI group (n = 74) utilized ChatGPT with carefully crafted prompts for writing feedback. A survey measuring emotional engagement-including emotional connections, sense of belonging, motivation and enthusiasm, and emotional reactions-was administered at the beginning and end of the semester. Mixed effects modelling was used for data analysis. Results indicated that all four types of emotional engagement were significantly higher in the GenAI group, with particularly pronounced outcomes in emotional connections, sense of belonging, and motivation and enthusiasm. These findings suggest that while GenAI feedback can be effectively integrated into EFL writing instruction, its combination with human interaction remains crucial. In particular, learners' emotional reaction to GenAI use is a key issue. The study concludes with pedagogical implications for leveraging GenAI in EFL writing contexts.
Prompt learning is an effective way to adapt pre-trained models to downstream tasks by training a small number of additional learnable prompts. Recent studies address several early challenges by combining generalized knowledge from frozen pre-trained VL models with task-specific knowledge from training data as guidance for prompt learning. However, existing methods still struggle with the generalization-adaptation (GA) trade-off dilemma: excessive reliance on generalized knowledge hinders adaptation to downstream tasks, while overemphasis on task-specific knowledge undermines the inherent generalization capabilities of pre-trained models. To address this issue, we propose a novel prompt learning method called Prompt Learning with Knowledge Regularization (PLKR). PLKR effectively mitigates the GA trade-off dilemma by offering greater flexibility in adapting to task-specific knowledge while minimizing the disruption of pre-trained knowledge. Specifically, we propose category-invariant and topology-invariant knowledge regularization to preserve generalized knowledge: the former enhances category-level discriminative capabilities while allowing flexible task-specific learning, and the latter maintains global topological stability during adaptation to new tasks. Through the proposed regularization, PLKR improves the performance on both base and new tasks. We evaluate the effectiveness of our approach on four representative tasks over 11 datasets. Experimental results show our method outperforms existing SOTA methods by a large margin.
