ACE Seminar: Efficient and Scalable Oblivious User Matching

Speaker: Dr Kasper Bonne Rasmussen

Date/Time: 10-Dec-2015, 16:00 UTC




Matching two or more users with related interests is an important and general primitive, applicable to a wide range of scenarios including job hunting, friend finding, and dating services. Existing on-line or off-line matching services requires participants to trust a third party server with their preferences. This raises security and privacy issues if the third party turns malicious or is compromised. This talk will introduce two privacy-preserving protocols: server-led matching and user-led matching. In the first protocol, potential matching pairs (e.g., users, companies) are selected by the server, which collects and combines each party's preference. In the second, entities are allowed to express their preference for any party---regardless of whether the other party is known to the server. With server-led matching, users reveal no information to the server; the server's role is simply to relay messages. In user-led matching, the server only learns which users match. The protocols are scalable, i.e., preferences can be matched in constant time, and formal security and functionality requirements will be presented. The protocols are presented as a framework for generic server-led and user-led matching, and we will cover the security proofs for two particular instantiations within this framework.



Dr Kasper Bonne Rasmussen personal page could be find here.



This page was last modified on 27 Mar 2014.