Recent Breakthroughs in Obfuscation


Date
Apr 1, 2016 12:00 AM

Who: Elena Pagnin When: Friday April 1, 11:00 Where: Room 8103 Title: {{ page.title }}

Abstract: This talk is supposed to give an overview of the state of the art in the area of Homomorphic Encryption (HE) and Multi-Linear Maps (MLM). The final section of the talk will deal with the definition and application of indistinguishable Obfuscation. More precisely, I will recall the syntax of FHE schemes and provide a trivial (extremely inefficient) construction of a FHE scheme. I will then cite the most significant papers in the area, highlight the trend and evolution of FHE schemes over the years (since Gentry’s first construction in 2009) and finally conclude with a slide that shows the timing required to perform homomorphic operations on a FHE scheme implemented by Halevi. The main focus of the talk is on Muliti-Linear Maps (MLM). Roughly speaking, MLMs can be thought of as a FHE scheme in which there is no decryption, but it is possible to check if two (top level) cipher-texts are equal or not. This latter property is called zero testing and it is the main feature, but also weakness, of MLMs. I will present the second MLM candidate, the CLT13, which is based on relatively easy mathematical tools (congruences and Chinese Remainder Theorem), and its cryptanalysis (zeroing attack). The talk will end with an introduction to obfuscation. >From an abstract point of view, one could think at obfuscation as a MLM without zero testing and in which only a pre-determined sequence of operations is allowed. In practice, an obfuscator is an algorithm that takes as input a program P and outputs another program O(P). such that O(P) and P have equal outputs on equal inputs and O(P) is an intelligible program.

Previous Talks

{: .t60 } {% include list-posts tag=‘csstalk’%}