Section: New Results

Results for Axis 3: Building a secure network stack

We designed new efficient protocol for privacy-preserving signed set intersection cardinality using blinded BLS signatures over bilinear maps and demonstrated its utility in machine learning for abuse detection in decentralised online social networks. The paper was presented at DPM 2016 [21].

We analyzed real-world mobility data based on cell tower traces, and illustrated how cell tower trace data can be used to identify patterns of life. We then used these results to predict future locations over a 24h period in 15 minute intervals with 80% accuracy [43].

We improved the Taler payment system protocol [13] to (1) reduce storage requirements for the exchange, which was the dominant cost, and (2) reduce security assumptions by avoiding the use of AES entirely.

We adapted the payment handshake to work even if JavaScript is disabled for the Web page, and adjusted the protocol to match discussions for future Web payment protocols from W3c. The protocol was extended with accounting functions to allow merchants to trace payments for their back office requirements. The user interface of the Taler wallet was streamlined, the wallet can finally get change, and the extension was made to work with Firefox. A public demonstrator was launched at https://demo.taler.net/.