Recently, some researches are devoted to the topic of end-to-end learning a physical layer secure communication system based on autoencoder under Gaussian wiretap channel. However, in those works, the reliability and security of the encoder model were learned through necessary decoding outputs of not only legitimate receiver but also the eavesdropper. In fact, the assumption of known eavesdropper's decoder or its output is not practical. To address this issue, in this paper we propose a dual mutual information neural estimation (MINE) based neural secure communications model. The security constraints of this method is constructed only with the input and output signal samples of the legal and eavesdropper channels and benefit that training the encoder is completely independent of the decoder. Moreover, since the design of secure coding does not rely on the eavesdropper's decoding results, the security performance would not be affected by the eavesdropper's decoding means. Numerical results show that the performance of our model is guaranteed whether the eavesdropper learns the decoder himself or uses the legal decoder.
Recently, some researches are devoted to the topic of end-to-end learning a physical layer secure communication system based on autoencoder under Gaussian wiretap channel. However, in those works, the reliability and security of the encoder model were learned through necessary decoding outputs of not only legitimate receiver but also the eavesdropper. In fact, the assumption of known eavesdropper's decoder or its output is not practical. To address this issue, in this paper we propose a dual mutual information neural estimation (MINE) based neural secure communications model. The security constraints of this method is constructed only with the input and output signal samples of the legal and eavesdropper channels and benefit that training the encoder is completely independent of the decoder. Moreover, since the design of secure coding does not rely on the eavesdropper's decoding results, the security performance would not be affected by the eavesdropper's decoding means. Numerical results show that the performance of our model is guaranteed whether the eavesdropper learns the decoder himself or uses the legal decoder.
Wireless communications are particularly vulnerable to eavesdropping attacks due to their broadcast nature. To effectively deal with eavesdroppers, existing security techniques usually require accurate channel state information (CSI), e.g., for friendly jamming (FJ), and/or additional computing resources at transceivers, e.g., cryptography-based solutions, which unfortunately may not be feasible in practice. This challenge is even more acute in low-end IoT devices. We thus introduce a novel deep learning-based FJ framework that can effectively defeat eavesdropping attacks with imperfect CSI and even without CSI of legitimate channels. In particular, we first develop an autoencoder-based communication architecture with FJ, namely, AEFJ, to jointly maximize the secrecy rate and minimize the block error rate (BLER) at the receiver without requiring perfect CSI of the legitimate channels. In addition, to deal with the case without CSI, we leverage the mutual information neural estimation (MINE) concept and design a MINE-based FJ scheme that can achieve comparable security performance to the conventional FJ methods that require perfect CSI. Extensive simulations in a multiple-input-multiple-output (MIMO) system demonstrate that our proposed solution can effectively deal with eavesdropping attacks in various settings. Moreover, the proposed framework can seamlessly integrate MIMO security and detection tasks into a unified end-to-end learning process. This integrated approach can significantly maximize the throughput and minimize the BLER, offering a good solution for enhancing communication security in wireless communication systems.
Practical codes for the Gaussian wiretap channel are designed aiming at satisfying information-theoretic metrics to ensure security against a passive eavesdropper (Eve). Specifically, a design criterion is introduced for the coset coding scheme in order to satisfy a strong secrecy condition described with the mutual information between the secret message and Eve's observation. In addition, mutual information neural estimation (MINE) powered from deep learning tools is applied in order to directly compute the information-theoretic security constraint, and verify the proposed solutions. It is shown that finite-length coset codes can indeed ensure secure transmission from an information-theoretic perspective.
To avoid the non-convex and complicated secrecy rate optimization in the multiple-input and multiple-output (MIMO) system, deep learning (DL) is introduced in the secure precoder design recently. However, existing works neglect the eavesdropper with learning capacity, which is constructed by neural networks (NNs) and capable of enhancing its receiver NN according to the optimal secure transmitter. In this letter, we propose an adversarial learning-based secure autoprecoder (ASAP) for the MIMO wiretap channels, where the modulation and the precoder are jointly optimized and adversarially trained with the eavesdropping receiver. Numerical results verify the convergence of the ASAP method, which enables an optimal trade-off between the information secrecy and reliability. Moreover, joint optimization is especially important against the evolved eavesdropper.
Mutual information neural estimation (MINE) and related variational estimators are increasingly used as empirical eavesdropper-leakage indicators in wireless physical-layer security (PLS) systems that lack a closed-form secrecy rate. We ask whether, given a matched estimation target, such estimators recover the true eavesdropper mutual information (MI). Across three wiretap systems of increasing complexity—a Gaussian wiretap, an orthogonal frequency-division multiple access (OFDMA) Rayleigh-fading wiretap, and a discrete-codebook sparse code multiple access (SCMA) wiretap test bed—we compare each held-out estimate against its own numerically computed reference, spanning four lower-bound estimators (DV/MINE, NWJ, SMILE, InfoNCE) and one upper-bound estimator (CLUB). Three findings emerge. i) A three-tier reliability hierarchy (estimator-vs-reference slopes 0.98/0.76/0.05 ): the estimate tracks the truth for the Gaussian wiretap, underestimates it for the discrete codebook, and collapses for the OFDMA fading channel. ii) The fading collapse persists even with the matched target (YE,HE) —recovering only 5–17% of the true conditional MI despite larger budgets and longer training, consistent with a function-class/finite-sample limitation under the tested protocols rather than a simple training-length failure. All four lower-bound estimators collapse; CLUB is not a reliable substitute—it overestimates the truth on the calibrated systems yet itself falls below the true MI under fading—so neither bound direction gives a reliable absolute leakage value. iii) Paired-design rankings are preserved only for single-parameter, large-separation comparisons; under multiple coupled variables ranking falls to chance on the hardest high-leakage pairs and is at best weakly rank-correlated overall (Kendall τ=0.27 ). We conclude that a near-zero neural MI estimate can coincide with substantial true leakage, so—for the tested estimators, critic families, sample budgets, and channel models—these estimators are unsafe as absolute eavesdropper-leakage indicators in the tested fading-mixture regime and should be used for comparative ranking only under matched, single-parameter, clearly separated designs, not for coupled multi-variable selection.
To achieve new applications for 5G communications, physical layer security has recently drawn significant attention. In a wiretap channel system, our goal is to minimize information leakage to an eavesdropper while maximizing the performance of transmission to the desired or legitimate receiver. Complicated systems or channel models make it difficult to design secrecy systems based on the information theory. In this paper, we propose a deep learning-based transceiver design for secrecy systems as an alternative. Specifically, we modify the loss function design of a variational autoencoder, which is a special type of neural network, making it possible to provide both robust data transmission and security in an unsupervised fashion. We further investigate the impact of an imperfect channel state information and use simulation results to prove that our approach can outperform the existing learning-based methods.
Secure communication under channel uncertainty is an important and challenging problem in physical-layer security and cryptography. In this dissertation, we take a \nfundamental information-theoretic view at three concrete settings and use them to shed insight into efficient secure communication techniques for different scenarios under channel uncertainty. \n \nFirst, a multi-input multi-output (MIMO) Gaussian broadcast channel with two receivers and two messages: a common message intended for both receivers (i.e., channel \nuncertainty for decoding the common message at the receivers) and a confidential message intended for one of the receivers but needing to be kept asymptotically perfectly secret from the other is considered. A matrix characterization of the secrecy capacity region is established via a channel-enhancement argument and an extremal entropy inequality previously established for characterizing the capacity region of a degraded compound MIMO Gaussian broadcast channel. \n \nSecond, a multilevel security wiretap channel where there is one possible realization for the legitimate receiver channel but multiple possible realizations for the eavesdropper channel (i.e., channel uncertainty at the eavesdropper) is considered. A coding scheme is designed such that the number of secure bits delivered to the legitimate receiver depends on the actual realization of the eavesdropper channel. More specifically, when the eavesdropper channel realization is weak, all bits delivered to the legitimate receiver need to be secure. In addition, when the eavesdropper channel realization is strong, a prescribed part of the bits needs to remain secure. We call such codes security embedding codes, referring to the fact that high-security bits are now embedded into the low-security ones. We show that the key to achieving efficient security embedding is to jointly encode the low-security and high-security bits. In particular, the low-security bits can be used as (part of) the transmitter randomness to protect the high-security ones. \n \nFinally, motivated by the recent interest in building secure, robust and efficient distributed information storage systems, the problem of secure symmetrical multilevel diversity coding (S-SMDC) is considered. This is a setting where there are channel uncertainties at both the legitimate receiver and the eavesdropper. The problem of encoding individual sources is first studied. A precise characterization of the entire admissible rate region is established via a connection to the problem of secure coding over a three-layer wiretap network and utilizing some basic polyhedral structure of the admissible rate region. Building on this result, it is then shown that the simple coding strategy of separately encoding individual sources at the encoders can achieve the minimum sum rate for the general S-SMDC problem.