| 摘要 |
The disclosed embodiments provide a method and system for performing regularized model adaptation for in-session recommendations. During operation, a server transmits to multiple clients a first global version of a statistical model for generating recommendations to users. At each client, during first user session with a user of the client, the first global version is used to output one or more recommendations to the user and, based on the user's feedback, updates to the global version are generated in real-time to create a first personalized version of the statistical model. At the end of the first user session, the updates to the first global version of the statistical model are transmitted to the server for use in producing a second global version of the statistical model by the server, which is again distributed to clients. |
