Cryptography in an Unbounded Computational Model

Yuval Ishai, Eyal Kushilevitz Z

Abstract We revisit the following open problem in information-theoretic cryptography: Does the communication complexity of unconditionally secure computation depend on the computational complexity of the function being computed? For instance, can computationally unbounded players compute an arbitrary function of their inputs with polynomial communication complexity and a linear threshold of unconditional privacy? Can this be done using a constant number of communication rounds? We provide an explanation for the difficulty of resolving these questions by showing that they are closely related to the problem of obtaining efficient protocols for (information-theoretic) private information retrieval and hence also to the problem of constructing short locally-decodable error-correcting codes. The latter is currently considered to be among the most intriguing open problems in complexity theory. Keywords. Information-theoretic cryptography, secure multiparty computation, private information retrieval, locally decodable codes. 1 Introduction In STOC 1990, Beaver, Micali, and Rogaway [5] posed the following question: Is there a constant-round protocol that allows k computationally unbounded players to defeat a computationally unbounded adversary, while using only a polynomial amount of communication in the total length of their inputs?

Eike Kiltz, Vinod Vaikuntanathan

Zvika Brakerski, Ran Canetti, Luowen Qian

In the classical model of computation, it is well established that one-way functions (OWF) are minimal for computational cryptography: They are essential for almost any cryptographic application that cannot be realized with respect to computationally unbounded adversaries. In the quantum setting, however, OWFs appear not to be essential (Kretschmer 2021; Ananth et al., Morimae and Yamakawa 2022), and the question of whether such a minimal primitive exists remains open. We consider EFI pairs - efficiently samplable, statistically far but computationally indistinguishable pairs of (mixed) quantum states. Building on the work of Yan (2022), which shows equivalence between EFI pairs and statistical commitment schemes, we show that EFI pairs are necessary for a large class of quantum-cryptographic applications. Specifically, we construct EFI pairs from minimalistic versions of commitments schemes, oblivious transfer, and general secure multiparty computation, as well as from QCZK proofs from essentially any non-trivial language. We also construct quantum computational zero knowledge (QCZK) proofs for all of QIP from any EFI pair. This suggests that, for much of quantum cryptography, EFI pairs play a similar role to that played by OWFs in the classical setting: they are simple to describe, essential, and also serve as a linchpin for demonstrating equivalence between primitives.

Elette Boyle, Mohammad Mahmoody

Michael O. Rabin, Yan Zong Ding

Traditional modern cryptography relies on unproven complexity assumptions that certain problems, such as integer factorization, are computationally hard, and the adversary is bounded in computing power. However, advances in algorithms and computing technology may render current cryptosystems insecure, and lead to the decryption of past secret messages. On the other hand, Shannon's pessimistic negative result essentially implies that if the adversary is all-powerful and has complete access to the communication line, then efficient practical solutions for information-theoretic security do not exist. This thesis is a study of cryptography in the bounded storage model, where provable everlasting security can be achieved efficiently without any complexity assumption. It first presents an efficient scheme for encryption and authentication that is provably secure against strong adaptive attacks, by an adversary who is computationally unbounded and only storage-bounded. A major new result is that the shared secret key employed by the sender and the receiver can be re-used to send an exponential number of messages. Our encryption scheme further enjoys the surprising property that even if the secret key is captured by the adversary after the transmission of the ciphertext, the secrecy of the message is assured. The above is joint work with Michael Rabin. Oblivious transfer is a cryptographic protocol that can serve as a basic primitive for all of cryptography. This thesis presents an implementation of oblivious transfer in the bounded storage model, which is again provably secure without complexity assumption.