Kyber is a KEM based their security on the Modular Learning with Errors problem and was selected in the second round of NIST Post-quantum standardization process. Before we put Kyber into practical application, it is very important to assess its security in hard practical conditions especially when the Fujisaki-Okamoto transformations are neglected. In this paper, we propose an efficient key mismatch attacks on Kyber, which can recover one participant's secret key if the public key is reused. We first define the oracles in which the adversary is able to launch the attacks. Then, we show that by accessing the oracle multiple times, the adversary is able to recover the coefficients in the secret key. Furthermore, we propose two strategies to reduce the queries and time in recovering the secret key. It turns out that it is actually much easier to use key mismatch attacks to break Kyber than NewHope, another NIST second round candidate, due to their different design structures. Our implementations have demonstrated the efficiency of the proposed attacks and verified our findings. Another interesting observation from the attack is that in the most powerful Kyber-1024, it is easier to recover each coefficient compared with that in Kyber-512 and Kyber-768. Specifically, for Kyber-512 on average we recover each coefficient with 2.7 queries, while in Kyber-1024 and 768, we only need 2.4 queries. This demonstrates further that implementations of LWE based schemes in practice is very delicate.

The aim of the present paper is to suggest that statistical physics provides the correct language to understand the practical behavior of the LLL algorithm, most of which are left unexplained to this day. To this end, we propose sandpile models that imitate LLL with compelling accuracy, and prove for these models some of the most desired statements regarding LLL. We also formulate a few conjectures that formally capture our heuristics and would serve as milestones for further development of the theory.

We use the signal function from RLWE key exchange to derive an efficient zero knowledge authentication protocol to validate an RLWE key p=as+e with secret s and error e in the Random Oracle Model (ROM). With this protocol, a verifier can validate that a key presented to him by a prover P is of the form p=as+e with s,e small and that the prover knows s. We accompany the description of the protocol with proof to show that it has negligible soundness and completeness error. The soundness of our protocol relies directly on the hardness of the RLWE problem. The protocol is applicable for both LWE and RLWE but we focus on the RLWE based protocol for efficiency and practicality. We also present a variant of the main protocol with a commitment scheme to avoid using the ROM.

We use an RLWE-based key exchange scheme to construct a simple and efficient post-quantum oblivious transfer based on the Ring Learning with Errors assumption. We prove that our protocol is secure in the Universal Composability framework against static malicious adversaries in the random oracle model. The main idea of the protocol is that the receiver and the sender interact using the RLWE-based key exchange in such a way that the sender computes two keys, one of them shared with the receiver. It is infeasible for the sender to know which is the shared key and for the receiver to get information about the other one. The sender encrypts each message with each key using a symmetric-key encryption scheme and the receiver can only decrypt one of the ciphertexts. The protocol is extremely efficient in terms of computational and communication complexity, and thus a strong candidate for post-quantum applications.

In this paper we show how to totally break the fully homomorphic encryption scheme of eprint 2017/458.