Special Session 183: Mathematics in Cryptography and Codes

Cryptanalytic Parameter Recovery of Neural Networks: Some Recent Advances

Yi Chen Institute of Advanced Study, Tsinghua University Peoples Rep of China Co-Author(s):

Abstract: The problem of model extraction in machine learning has been studied for over thirty years. Its most challenging goal --- functionally equivalent extraction in the black-box setting --- is achieved via parameter recovery. Since Crypto 2020, researchers have made significant progress by approaching it through the lens of cryptanalysis. In this talk, we will briefly review this problem and introduce some recent results in this emerging direction, with a special focus on attacks in the hard-label setting.

Secure Network Function Computation

Xuan Guang Nankai University Peoples Rep of China Co-Author(s):    Xuan Guang, Yang Bai, Raymond W. Yeung

Abstract: to be updated

Secure Network Function Computation for Linear Functions

Xuan Guang Nankai University Peoples Rep of China Co-Author(s):

Abstract: We will introduce the problem of (information-theoretically) secure network function computation. For the problem, a target function, of which the inputs are generated at multiple source nodes, is required to be computed with zero error at a sink node over a network, while a wiretapper, who can access any one but not more than one wiretap set in a given collection of wiretap sets, is not allowed to obtain any information about a security function of the source messages. We are interested in characterizing the secure computing capacity for this problem, which is defined as the maximum average number of times the target function can be securely computed with zero error for one use of the network. In general, the characterization of this secure capacity with this general setup is overwhelmingly difficult. In this talk, we focus on the secure model for linear functions with the wiretapper being able to eavesdrop any subset of edges in the network up to a certain size, referred to as the security level.

Security Analysis of SHA-3 Hash Functions

Meicheng Liu Institute of Information Engineering, Chinese Academy of Sciences Peoples Rep of China Co-Author(s):

Abstract: The cryptographic hash algorithm plays a pivotal role in modern cryptography. It compresses messages of arbitrary length into fixed-length digests, which are utilized for data integrity protection, digital signatures, identity authentication, password protection protocols, electronic payment protocols, blockchain and so on. The SHA-3 (Secure Hash Algorithm 3) family of hash algorithms, the third-generation cryptographic hash standard developed by the National Institute of Standards and Technology (NIST), has been extensively studied in the cryptography community over the past decade. This report will summarize the latest research progress on the security analysis of SHA-3 hash algorithms.

A New Configuration for 3rd-FHE Bootstrapping from the Key Switching

Han Wang Institute of Information Engineering, CAS Peoples Rep of China Co-Author(s):

Abstract: This work introduces a new configuration of the $GSW$ FHE (Gentry, Sahai, Waters~Crypto 2013), with a squared gadget ,batching and scale-based homomorphic operation, from the key switching operations. This configuration offers improved efficiency compared to existing approaches. By utilizing our proposed method as the underlying building block, we can accelerate 3rd-FHE bootstrapping implementations, including the libraries of $OpenFHE$ and $TFHE$. We conduct comprehensive experiments to evaluate the concrete performance of our method, demonstrating improvements of more than 2 times faster.

New Results on Elliptic Curve Hidden Number Problem for ECDH Key Exchange

Jun Xu Institute of Information engineering, CAS Peoples Rep of China Co-Author(s):    Jun Xu, Santanu Sarkar, Huaxiong Wang, Lei Hu

Abstract: The Elliptic Curve Hidden Number Problem (EC-HNP) was introduced at Asiacrypt 2001, and its Diffie-Hellman variant proposed at PKC 2017 serves for evaluating ECDH bit security and side-channel cryptanalysis. In this talk, we present the Coppersmith method for solving the multivariate modular polynomials in this variant. For a fixed elliptic curve over $\mathbb{F}_p$ with large prime $p$, if an oracle returns $\delta/\log_2 p$ fraction of the most (least) significant bits of the ECDH key`s $x$-coordinate, we give a heuristic polynomial-time algorithm in $\log_2 p$ to recover the full secret. Our result allows any constant ratio in $(0,1)$, which improves over the bounds $5/6$ and $1/2$ in earlier works. Although no rigorous ECDH bit security is derived, experiments on NIST curves with small lattices validate the heuristic effectiveness.

A Simple Introduction to Secure Multi-Party Computation

Cong Zhang Tsinghua University Peoples Rep of China Co-Author(s):

Abstract: This talk will provide a concise introduction to secure multi-party computation (MPC), covering its fundamental concepts, including the definition of MPC, oblivious transfer (OT), and constructions based on garbled circuits.

Applications of Gaussian Sampling in Cryptography

Shiduo Zhang Tsinghua University Peoples Rep of China Co-Author(s):

Abstract: This presentation will systematically explain the application of Gaussian sampling in cryptographic design and the related parameter settings.