To improve our previous project titled as, "Privacy Preserving Information Retrieval", we are proposing a new client-centered approach based on language model. In this work, we propose a hierarchical language model based approach for privacy preserving search which attempts to create more plausible cover queries and handle sequentially edited user queries during a search session. Besides we have introduced a better technique to approximate query specificity which maintains same specificity level for cover queries compared to original query. Moreover, since this approach works in a hierarchical fashion, it is more effective in capturing user intent and using it during cover query generation which improves their plausibility.
myd1 / hierarchical-privacy-preserving-ir Goto Github PK
View Code? Open in Web Editor NEWThis project forked from zn8ae/hierarchical-privacy-preserving-ir
Hierarchical language model based approach for privacy preserving personalized web search
License: MIT License