Search

The recommendation system is widely used in the digital economy to establish relationship between users and products. However, the data is usually scattered in different organizations in the form of "data islands" to protect privacy. To solve the above problem, we proposed an efficient and secure recommendation system based on federated matrix factorization, in which every user trains the model locally with their own data and uploads the calculated gradient vector instead of the original data. In the scheme, one server verifies the aggregation results transmitted by other servers and completes the model update without revealing user privacy. Additionally, users are allowed to drop out during the training process, and each participant's behavior will be recorded to calculate the rewards or punishments. Finally, we evaluated the performance from different perspectives and proved its feasibility. [ABSTRACT FROM AUTHOR]

Checking the design specification for contradictions at the early phase of the software development process is crucial to ensure that the design is implementable. However, the high expressivity of OCL makes manual inconsistency checking a difficult task. In addition, the developers cannot detect these problems by OCL itself due to its lack of automated reasoning support. We investigate an approach to translating OCL invariants into OWL 2 DL axioms. We do this where the OCL expression contained in an invariant is converted to the corresponding OWL 2 DL class expression in a compositional way. Our approach covers the OCL expressions including four: PrimaryExp, RelationalExp, LogicalExp and IfExp types. Considering the distinction between the CWA and OWA, we achieve correct translation from RelationalExp using closure axiom. Also, we employ an ontology design pattern to overcome the limitations of OWL 2 DL expressivity when translating IfExp. Then inconsistency checking is done through description logic reasoning by the OWL 2 DL high-performance reasoner. We construct an inductive proof to establish the correctness of our translation approach. Moreover, we evaluate our approach using the implemented TUCO tool prototype. [ABSTRACT FROM AUTHOR]