Non-Malleable Codes for Bounded Depth, Bounded Fan-in Circuits
Non-malleable codes are a relaxation of error correcting codes, for settings in which privacy, but not necessarily correctness, is desired. Instead of requiring that after modification—i.e. tampering—of the codeword, the original message can always be recovered, non-malleable codes allow a different message to be recovered, as long as the recovered message is unrelated to the original message. This relaxation potentially allows for the construction of coding schemes for rich classes of tampering classes, beyond what can be done for error correcting codes. In applications, non-malleable codes are used to encode the memory of a device, and thus protect against (certain classes) of adversarial tampering.
Lower-Bounds on Assumptions behind Indistinguishability Obfuscation
In this talk, we first show that basing IO on a variety of assumptions (e.g., trapdoor permutations, bi-linear maps, etc) in a weakly black-box way is as hard as basing public-key encryption on one-way functions (in a non-black-box way). The latter has remained as one of the most challenging open questions in cryptography. Then, by combining our results with a recent result of Brakerski, Brzuska, and Fleischhacker, we rule out any fully black-box construction of IO from the same set of primitives assuming the existence of one-way functions and that the polynomial-hierarchy does not collapse.
New Inference Attacks on Order-Preserving and Order-Revealing Encryption
Accessing Data while Preserving Privacy
We present a new model of differentially private storage where differential privacy is preserved even against an attacker that controls the data and the queries made to it. We give a generic construction of differentially private storage that combines ORAM and differentially private sanitizers. We also provide efficient constructions and lower bounds for some specific query sets. We have implemented some of our algorithms, and report on their efficiency.Joint work with Georgios Kellaris, George Kollios, Kobbi Nissim, and Adam O’Neill.
Bio: Georgios is currently a Post-Doctoral Fellow at CRCS, Harvard University, and at Boston University. He received his Ph.D. degree in Computer Science and Engineering from the Hong Kong University of Science and Technology (2015), under the supervision of prof. Dimitris Papadias, and with the support of the Hong Kong Ph.D. Fellowship Scheme. He holds a 4-year B.Sc. in Informatics and Telecommunications from the University of Athens (2006) and a 2-year M.Sc. degree in Digital Systems from the University of Piraeus (2008). He has worked as a researcher at the University of Piraeus in Greece, the Singapore Management University and the Nanyang Technological University in Singapore, and at Boston University. His research interests include databases and differential privacy.