Motor-skill learning systems in XR rely on persistent cues. However, constant cueing can induce overreliance and erode memorization and skill transfer. We introduce a skill-adaptive, dynamically transparent ghost instructor whose opacity adapts in real time to learner performance. In a first-person perspective, users observe a ghost hand executing piano fingering with either a static or a performance-adaptive transparency in a VR piano training application. We conducted a within-subjects study (N=30), where learners practiced with traditional Static (fixed-transparency) and our proposed Dynamic (performance-adaptive) modes and were tested without guidance immediately and after a 10-minute retention interval. Relative to Static, the Dynamic mode yielded higher pitch and fingering accuracy and limited error increases, with comparable timing. These findings suggest that adaptive transparency helps learners internalize fingerings more effectively, reducing dependency on external cues and improving short-term skill retention within immersive learning environments. We discuss design implications for motor-skill learning and outline directions for extending this approach to longer-term retention and more complex tasks.

With growing environmental concerns and the push for sustainable urban development, promoting green travel has become a critical initiative. Urban transit systems face the challenge of integrating green initiatives with efficient transport routes, while sophisticated graph modeling enhances travel efficiency. However, blending historical and contemporary elements introduces complex variations in traffic networks, complicating feature extraction and clustering for information retrieval due to multi-scale spatial heterogeneity. Traditional methods often overlook key nuances by oversimplifying data relationships. We proposed M3 and validated the integration of GIS-based Attention-Cluster-GCN with Dueling Double Deep Q Network across various cities, enhancing urban travel with detailed information on green attraction recommendations, considering the usage of Multi-Purpose and Multi-Stakeholder for Multi-Scale Spatial Heterogeneity scenarios. Utilizing Attention-Based Reinforcement Graph Clustering refines modeling and emphasizes vital connections, enhancing personalized recommendation precision and clustering performance. Our method surpasses both conventional and advanced GNN methods, even in graph convolution-based deep reinforcement learning, achieving superior cluster separation and accuracy. Our sampling and ablation studies confirm the pivotal role of the attention mechanism and multi-scale features, showing a significant performance decline without attention. Our findings underscore the potential of graph clustering in making public transport more engaging and aligned with green attractions policies by recommendations, even amidst significant spatial heterogeneity.