Multi-Group ObScure Logging (MG-OSLo) A Privacy-Preserving Protocol for Private Web Search

The Web Search Engine (WSE) is a software system used to retrieve data from the web successfully. WSE uses the user's search queries to build the user's profile and provide personalized results. Users' search queries hold identifiable information that could compromise the privacy of the respective user. This work proposes a multi-group distributed privacy-preserving protocol (MG-OSLo) and tries to investigate the state-of-the-art distributed privacy-preserving protocols for computing web search privacy. The MG-OSLo comprises multiple groups in which each group has a fixed number of users. The MG-OSLo measures the impact of the multi-group on the user's privacy. The primary objective of this work is to assess local privacy and profile privacy. It aims at evaluating the impact of group size and group count on a user's privacy. Two grouping approaches are used to group the users in MG-OSLo, i.e. a non-overlapping group design and overlapping group design. The local privacy results reveal that the probability of linking a query to the user depends on the group size and group count. The higher the group size or group count, the lower the likelihood of relating the query to the user. The profile privacy computes the profile obfuscation level using a privacy metric Profile Exposure Level (PEL). Different experiments have been performed to compute the profile privacy of the subset of an AOL query log for two situations: i) self-query submissions allowed and ii) self-query submissions not allowed. The privacy achieved by MG-OSLo is compared with the modern privacy-preserving protocol UUP(e), OSLo, and Co-utile protocols. The results show that the MG-OSLo provided better results as compared to OSLo, UUP, and Co-utile. Similarly, the multi-group has a positive impact on local privacy and user profile privacy.

Automated summary

The paper introduces a multi-group distributed privacy-preserving protocol (MG-OSLo) to evaluate the impact of group size and count on user privacy. Experimental results show that MG-OSLo outperforms other protocols in user privacy protection.