Abstract:

In this talk, we will cover some exciting progress on the problem of Indistinguishability Obfuscation (proposed by Barak et. al. 2001). In a nutshell, an Indistinguishability Obfuscation scheme is an efficient compiler that takes as input a program and outputs a new program with the same input-behavior and only a polynomial slowdown, but in addition, we have the guarantee that the new program reveals minimal information about the original program. 

If realized securely and efficiently, such an obfuscation scheme would have huge consequences to both theory and practice. However, until now we did not know if it exists under any reasonably well-believed conjecture. Our work places iO onto “terra-firma”, by giving a construction that is as secure as several well-studied mathematical problems that are widely believed to be extremely hard to solve.  

In this talk, we will hear about indistinguishability obfuscation, why it is useful, how it can be constructed, and future work.