Shuffle Model

Cryptography and privacy in the shuffle model, where many parties send anonymous messages that reach the server in a uniformly random order.

Status	Statement	Tags
	Perfectly Correct Statistical ARE Statistical security with statistical correctness is settled by the source for every finite function. Perfect correctness is known in the computational setting and, in the statistical setting, only in a relaxed Las Vegas form where the evaluator may declare failure. 4 open	Additive Randomized Encodings Shuffle Modelcharacterizationadaptation (ai)
	Polylog Shuffle PIR, Negligible Error Settled in three neighbouring corners: O(n^gamma) communication with inverse-polynomial error and polynomially many queries; polylogarithmic communication with n^{O(log n)} queries; and negligible error with O(n/log n) communication. The polylogarithmic-and-negligible corner with polynomially many queries is open, and is ruled out for inner-outer constructions with an additive inner layer by the source’s own Theorem 6.5. 5 open	Private Information Retrieval Shuffle Modelcharacterizationadaptation (ai)
	MDSD Linear-Test Bias The source proves a non-tight reduction from standard decisional syndrome decoding and separately conjectures that the DOOM algorithm is the best attack; the linear-test bias is the one quantity it states it cannot bound. 5 open	Learning Parity With Noise Shuffle Model Syndrome Decodinglower-bound