Recently, recommendation systems have actively integrated Transformers to capture real-time context. However, these systems often suffer from generalization imbalance, where predictions are biased toward popular (head) items due to the sparsity and volatility inherent in session-based data. To address this challenge, this paper proposes MoE-SLMRec, a Mixture-of-Experts (MoE)-based recommendation model that selects expert networks based on session-level contextual information. The proposed model extracts a session latent representation, $h$, through a session-aware controller and forms balanced predictive characteristics across the entire data distribution via dynamic routing. Experimental results demonstrate that MoE-SLMRec significantly outperforms the baseline SLMRec, improving accuracy by 1.51 percentage points (from 18.76% to 20.27%). Furthermore, the model achieved state-of-the-art performance in Recall@20 (0.8358) and MRR@20 (0.3455), validating simultaneous improvements in both retrieval capability and ranking quality. Notably, the model effectively stabilized the performance for head items while coordinating the generalization trade-off between head and tail segments. By ensuring a favorable capacity–cost trade-off while maintaining robust performance, this study presents a promising alternative under session-based recommendation settings, facilitating scalable deployment in real-time recommendation services.

This paper presents an approach that integrates tiered storage into the Log-Structured Merge (LSM)-tree to balance Key–Value Store (KVS) performance and storage financial cost trade-offs. The implementation focuses on applying tiered storage to LSM-tree-based KVS architectures, using both vertical and horizontal storage alignment strategies or a combination of both. Additionally, these configurations leverage key–value (KV) separation to further improve performance. Our experiments reveal that this approach reduces storage financial costs while offering trade-offs in write and read performance. For write-intensive workloads, our approach achieves competitive performance compared to a fast NVMe Solid State Drive (SSD)-only approach while storing 96% of data on more affordable SATA SSDs. Additionally, it exhibits lookup performance comparable to BlobDB, and improves range query performance by 1.8x over RocksDB on NVMe SSDs. Overall, the approach results in a 49.5% reduction in storage financial cost compared to RocksDB and BlobDB on NVMe SSDs. The integration of selective KV separation further advances these improvements, setting the stage for future research into offloading remote data in LSM-tree tiered storage systems.

This study proposes a Personalized Deep Learning-Based Sleep Recommendation System Using Lifelog Data (PDSRS-LD). Traditional sleep research primarily relies on bio signals such as EEG and ECG recorded during sleep but often fails to sufficiently reflect the influence of daily activities on sleep quality. To address this limitation, we collect lifelog data such as stress levels, fatigue, and sleep satisfaction via wearable devices and use them to construct individual user profiles. Subsequently, real sleep data obtained from an AI-powered motion bed are incorporated for secondary training to enhance recommendation performance. PDSRS-LD considers comprehensive user data, including gender, age, and physical activity, to analyze the relationships among sleep quality, stress, and fatigue. Based on this analysis, the system provides personalized sleep improvement strategies. Experimental results demonstrate that the proposed system outperforms existing models in terms of F1 score and Average Precision (mAP). These results suggest that PDSRS-LD is effective for real-time, user-centric sleep management and holds significant potential for integration into future smart healthcare systems.

Wearable lifelog data provides continuous physiological and behavioral signals, enabling contextual understanding of users’ real-world states. However, conventional recommendation models rely heavily on static behavioral logs and fail to incorporate contextual variability, leading to limited personalization and biased decision-making. To address these limitations, this study proposes a case-based Mixture-of-Experts (MoE) framework that integrates case routing, federated learning (FL), and reinforcement learning (RL) for context-adaptive action recommendation. The model first classifies user states—derived from multimodal lifelog features such as heart-rate variability, activity intensity, sleep patterns, and fatigue—into multiple context-specific cases. A shared DeepFM–SASRec backbone encodes both static and sequential physiological patterns, while case-specific Adapter, LoRA, and MoE expert heads refine representations to capture intra-case variability. Federated learning enables robust model training under heterogeneous; privacy-sensitive lifelog distributions, and reinforcement learning optimizes long-term physiological improvement by modeling state transitions and action outcomes. Experimental results demonstrate that the proposed framework achieves substantial gains in both accuracy and fairness. The final variant improves HitRate@20 from 0.628 to 0.881 (+40.3%) and reduces Bias Ratio from 0.92 to 0.64 (–30%), effectively balancing precision and behavioral diversity. These findings confirm that combining case-based routing, MoE specialization, FL-driven robustness, and RL-based policy optimization provides a powerful and scalable foundation for next-generation lifelog-driven action recommendation systems.