The field of predictive modeling and anomaly detection is experiencing significant growth, driven by the increasing availability of large datasets and advances in machine learning techniques. Researchers are exploring new approaches to improve the accuracy and robustness of predictive models, particularly in complex and dynamic environments. One notable trend is the integration of multiple modeling techniques, such as ensemble methods and hybrid representation frameworks, to capture diverse patterns and relationships in data. Additionally, there is a growing focus on developing methods that can adapt to changing conditions and detect anomalies in real-time, with applications in areas such as cloud computing, smart homes, and urban pollution monitoring. Noteworthy papers include:

• Set-Valued Transformer Network for High-Emission Mobile Source Identification, which proposes a novel approach to identifying high-emission vehicles using a set-valued transformer network.
• Online Ensemble Transformer for Accurate Cloud Workload Forecasting in Predictive Auto-Scaling, which presents an online ensemble model for workload forecasting in cloud computing environments.
• HRS: Hybrid Representation Framework with Scheduling Awareness for Time Series Forecasting in Crowdsourced Cloud-Edge Platforms, which introduces a hybrid representation framework for time series forecasting in cloud-edge platforms.
• Artificial Intelligence-Based Multiscale Temporal Modeling for Anomaly Detection in Cloud Services, which proposes a multiscale temporal modeling approach for anomaly detection in cloud services.
• DualNILM: Energy Injection Identification Enabled Disaggregation with Deep Multi-Task Learning, which presents a deep multi-task learning framework for non-intrusive load monitoring in smart homes and building applications.