Modeling feature interactions for context-aware QoS prediction of IoT services

QoS prediction in intelligent edge computing based on feature learning

“…Meanwhile, the above DL based methods do not pay much attention to the effect of higher order interaction of features. To enrich the interacting manner between features of users and service, Chen et al [26] proposed a context-aware feature interaction modeling (CFM) to capture both memorization and generalization by jointly considering low-order feature interactions with FM and high-order feature interactions with a MLP and deep cross network(DCN). Wang et al [23] proposed LAFIL leverage Compressed Interaction Network (CIN) for feature interaction learning of the location information only.…”

“…Wu et al [27] proposed a universal deep neural model(DNM) for QoS prediction with contexts, which respectively learning the features interaction from user-side and service-side, but the features interaction between the user and the service are ignored. Unfortunately, existing feature interaction learning methods [23,26,27] based on DL do not distinguish the importance of different feature interactions. They assign the same weight to all feature interactions, while the identification and process of low correlation feature interactions are ignored, seriously affecting the accuracy of QoS prediction.…”

“…Still, the contextual information is not fully exploited and utilized effectively in the interaction between the user and the service. 2) Approaches based on feature learning [26,27] typically focus only on the interaction of explicit features. However, the implicit features are of vital in QoS prediction, which is usually ignored by these methods in their interaction layer.…”

Dynamic Context-Aware Workflow Management Architecture for Efficient Manufacturing: A Ros-Based Case Study

“…Context-awareness introduces new capabilities and opportunities for improving efficiency, adaptability, and responsiveness in manufacturing processes [10]. Context-aware workflow management systems can dynamically adapt and optimize processes by analyzing real-time information and performing intelligent decision-making [11]. These systems leverage contextual factors such as resource availability, resource status, and environmental changes [12] to make decisions on dynamically adjusting the workflow during execution.…”