Friday, March 11 (In Person!)
We are very stimulate to bring the Boston-area crypto community back together for another in-person crypto sidereal day with four exciting talks .
When : *IN-PERSON* Crypto Day on Friday, March 11 at BU .
Where: BU ’ s Barrister Hall ( located at the ground floor of the law school tugboat [ map ] ) .
Logistics: All attendants are required to be in full vaccinated, and wear masks when indoors at all times.
Reading: Friday, March 11 (In Person!)
We look ahead to seeing you all .
Organizers: Ran Canetti, Yael Kalai, Vinod Vaikuntanathan and Daniel Wichs .
|9:00 – 9:30||Welcome and Coffee|
|9:30 – 10:30||Yevgeniy Dodis, NYU
|11:00 – 12:00||Guy Bresler, MIT
Average-case Reductions amongst Statistical Problems
|12 – 1:30||Lunch (provided)|
|1:30 – 2:30||David Heath, Georgia Tech
EpiGRAM: Practical Garbled RAM
|2:30 – 3:00||Coffee|
|3:00 – 4:00||Gabe Kaptchuk, Boston University
Weaving Social Accountability into Cryptographic Protocols
Speaker: Yevgeniy Dodis
Title: Small-Box Cryptography
One of the ultimate goals of symmetric-key cryptography is to find a rigorous theoretical framework for building complex cryptanalytic primitives from little components, such as cryptanalytic S-boxes. unfortunately, a fundamental obstacle towards reaching this finish comes from the fact that traditional security proof can not get security beyond 1/2^n, where n is the size of the corresponding belittled component. As a result, anterior demonstrably secure approaches — which we call “ big-box cryptography ” — always made n larger than the security parameter, which led to several problems. Most notably, the design was besides coarse to very explain virtual constructions, as ( arguably ) the most concern design choices happening when instantiating such “ big-boxes ” were completely abstracted out .
In this ferment, we introduce a novel paradigm ( or heuristic ? ) for building bad objects from small objects, which we call small-box cryptanalysis .
As an exemplification, we apply our framework to the analysis of SPN ciphers ( e.g, generalizations of AES ), getting quite fair concrete severity estimates for the resulting ciphers. We besides apply our framework to the design of current ciphers. here, however, we focus on the simplicity of the resulting construction, for which we besides managed to find a directly “ big-box ” -style security justification, under a well-studied and wide believed exact Linear Parity with Noise ( XLPN ) assumption .
overall, we hope that our work might initiate newfangled follow-up results in the sphere of small-box cryptanalysis .
joint oeuvre with Harish Karthikeyan and Daniel Wichs .
Speaker: Guy Bresler
Title: Average-Case Reductions Amongst Statistical Problems
In this speak I will describe a few childlike average-case reduction techniques and use these techniques to show how the computational phase transitions in a variety of statistical problems with widely varying structures all follow from a slight abstraction of the plant clique speculation. Some of these problems are robust sparse analogue arrested development, tensor PCA, and certain dense stochastic engine block models. The talk is based on joint work with Matthew Brennan ( hypertext transfer protocol : //arxiv.org/abs/2005.08099 ) .
Speaker: David Heath
Title: EpiGRAM: Practical Garbled RAM
Yao ’ south Garbled Circuit ( GC ) is a foundational technique for achieving dependable multiparty calculation ( MPC ). GC allows two parties to securely compute any Boolean racing circuit. however, many computations are naturally expressed as RAM programs, not as Boolean circuits. While any RAM program can be polynomially reduced to a circuit, in practice the decrease is often prohibitively expensive .
Garbled RAM ( GRAM ) is a GC improvement that immediately supports random access arrays without adding extra rounds to the GC protocol. Hence, GRAM allows us to securely and directly compute RAM programs. The first non-trivial GRAM ( Lu and Ostrovsky, Eurocrypt 2013 ) required perennial evaluation of a PRF inside GC. unfortunately, this non-black-box use of a cryptanalytic primitive is enormously expensive. While several subsequent works improved diverse crucial aspects of GRAM, no works addressed the cardinal offspring : GRAM remained expensive .
In this talk, I will present EpiGRAM, a newfangled construction that dramatically improves GRAM ’ randomness efficiency. EpiGRAM improves over anterior work by three to four orders of magnitude and opens the door to using GRAM in implementations. Our key insight is a modern garbled datum structure that we call a lazy permutation network. This data social organization allows the falsify tour to efficiently ( and without supernumerary rounds ) outsource RAM accesses to the GC evaluator. In this talk, I will present the technical issues built-in to GRAM, and I will show how the faineant permutation network efficiently solves these problems .
EpiGRAM is joint work with Vladimir Kolesnikov ( Georgia Tech ) and Rafail Ostrovsky ( UCLA ) and will appear at Eurocrypt 2022 .
Speaker: Gabe Kaptchuk
Title: Weaving Social Accountability into Cryptographic Protocols
When deployed, cryptanalytic systems transform from strictly technical system into sociotechnical constructs. As a resultant role, cryptographers must take particular concern that it is potential to keep those who use these cryptanalytic systems maliciously can be avail accountable, by making pervert difficult and detectable. I investigate this trouble in the set of throughout code systems. First, I will discuss Abuse Resistant Law Enforcement Access Systems [ Eurocrypt21 ], which studies the problem of constructing law enforcement access systems ( i. politics backdoors ) that ensure accountability mechanisms in the case of unauthorized surveillance. Second, I will discuss Apple ’ s holocene CSAM scanning proposal, analyzing the organization through the lens of accountability .
Read more: A Few Thoughts on Cryptographic Engineering