Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Log inSign up
Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Blog
Job Board
Q3AS

© 2026 Aqora Quantum S.A.S.

TermsPrivacyLegal Notice

arXiv:2508.05176 · IEEE Globecom Workshops (GC Wkshps)

Neural Estimation of Information Leakage for Secure Communication System Design

Darius S. Heerklotz, Ingo Schröder, Pin-Hsun Lin, Christian Deppe, Eduard Axel Jorswieck

Published Dec 8, 2025• 4 citations
arXivHTMLPDF

Abstract

Underestimating the leakage can compromise physical layer security (PLS), while overestimating it may lead to inefficient system design. Therefore, a reliable leakage estimator is essential. Neural network-based estimators provide a data-driven way to estimate mutual information (M.I.) without requiring full knowledge of the channel or source distributions. In this work, we aim to scale the blocklength of a wiretap code such that the estimator can still feasibly operate. We propose an improved M.I. estimator based on the variational contrastive log-ratio upper bound (vCLUB) framework, tailored for both discrete and continuous variables. By using a mixture of Bernoulli experts parameterized by neural networks, the estimator is able to quantify information leakage in communication systems, which employ complex data processing like universal hash families (UHF). We further propose a method to utilize the proposed estimator to design the UHF for a wiretap code or secret key generation design. Simulation results show that prior methods significantly underestimate the M.I., particularly when using UHF for higher blocklengths (n ≫16). The proposed method can scale the blocklength up to 255, and we conjecture that the design can scale well to even higher blocklengths given adequate training data and model size. Additionally, we contend that our proposed estimator and adaptive hash design framework offer a practical approach for extending PLS considerations for wiretap channels into the finite blocklength regime.

DOI
10.1109/gcwkshps68340.2025.11591013

Related papers

Reliability and Collapse of Neural Mutual-Information Estimators for Eavesdropper-Leakage Assessment in Wireless Physical-Layer Security

Gang Liu, Guosheng Xu, Chenyu Wang +1 more

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 $(Y_{E},H_{E})$ —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 $\tau = 0.27$ ). 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.

IEEE Open Journal of the Communications Society•Jan 1, 2026•0 citations

Performance Analysis of Finite Blocklength Transmissions Over Wiretap Fading Channels: An Average Information Leakage Perspective

Milad Tatar Mamaghani, Xiangyun Zhou, Nan Yang +2 more

Physical-layer security (PLS) is a promising technique to complement more traditional means of communication security in beyond-5G wireless networks. However, studies of PLS are often based on ideal assumptions such as infinite coding blocklengths or perfect knowledge of the wiretap link’s channel state information (CSI). In this work, we study the performance of finite blocklength (FBL) transmissions using a new secrecy metric — the average information leakage (AIL). We evaluate the exact and approximate AIL with Gaussian signaling and arbitrary fading channels, assuming that the eavesdropper’s instantaneous CSI is unknown. We then conduct case studies that use artificial noise (AN) beamforming to analyze the AIL in both Rayleigh and Rician fading channels. The accuracy of the analytical expressions is verified through extensive simulations, and various insights regarding the impact of key system parameters on the AIL are obtained. Particularly, our results reveal that allowing a small level of AIL can potentially lead to significant reliability enhancements. To improve the system performance, we formulate and solve an average secrecy throughput (AST) optimization problem via both non-adaptive and adaptive design strategies. Our findings highlight the significance of blocklength design and AN power allocation, as well as the impact of their trade-off on the AST.

arXiv:2308.13184•IEEE Transactions on Wireless Communications•May 21, 2024•11 citations

A Data-Driven Analysis of Secret Key Rate for Physical Layer Secret Key Generation From Wireless Channels

Toshiki Matsumine, Hideki Ochiai, Junji Shikata

This letter considers a data-driven approach to a secret key rate analysis of physical layer secret key generation based on mutual information neural estimator (MINE), which achieves the state-of-the-art estimation accuracy. Since MINE does not require any statistical knowledge of randomness sources, the proposed approach is applicable to any source model and even real data which cannot be fully expressed by a simple model. We first demonstrate by simulations that the proposed estimation provides good approximations to the theoretical secret key rate for a simple model, and then show that the achievable secret key rate highly depends on sources of randomness and employed channel estimation schemes for a realistic wireless channel model. As an application of our analysis, we present a method to train quantization through backpropagation that maximizes the estimated secret key rate.

IEEE Communications Letters•Oct 12, 2023•14 citations

On the Information Leakage Performance of Secure Finite Blocklength Transmissions over Rayleigh Fading Channels

Milad Tatar Mamaghani, Xiangyun Zhou, Nan Yang +2 more

This paper presents a secrecy performance study of a wiretap communication system with finite blocklength (FBL) transmissions over Rayleigh fading channels, based on the definition of an average information leakage (AIL) metric. We evaluate the exact and closed-form approximate AIL performance, assuming that only statistical channel state information (CSI) of the eavesdropping link is available. Then, we reveal an inherent statistical relationship between the AIL metric in the FBL regime and the commonly-used secrecy outage probability in conventional infinite blocklength communications. Aiming to improve the secure communication performance of the considered system, we formulate a blocklength optimization problem and solve it via a low-complexity approach. Next, we present numerical results to verify our analytical findings and provide various important insights into the impacts of system parameters on the AIL. Specifically, our results indicate that i) compromising a small amount of AIL can lead to significant reliability improvements, and ii) the AIL experiences a secrecy floor in the high signal-to-noise ratio regime.

arXiv:2401.11219•International Conference on Communications•Jun 9, 2024•3 citations

A Variational Autoencoder-Based Secure Transceiver Design Using Deep Learning

Chia-Hung Lin, Chao‐Chin Wu, Kuan‐Fu Chen +1 more

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.

Globecom•Dec 1, 2020•11 citations

Dual MINE-based Neural Secure Communications under Gaussian Wiretap Channel

Li, Jingjing, Zhuo Sun, Lei Zhang +1 more

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.

arXiv:2102.12918•Feb 25, 2021•1 citation
Underestimating the leakage can compromise physical layer security (PLS), while overestimating it may lead to inefficient system design. Therefore, a reliable leakage estimator is essential. Neural network-based estimators provide a data-driven way to estimate mutual information (M.I.) without requiring full knowledge of the channel or source distributions. In this work, we aim to scale the blocklength of a wiretap code such that the estimator can still feasibly operate. We propose an improved M.I. estimator based on the variational contrastive log-ratio upper bound (vCLUB) framework, tailored for both discrete and continuous variables. By using a mixture of Bernoulli experts parameterized by neural networks, the estimator is able to quantify information leakage in communication systems, which employ complex data processing like universal hash families (UHF). We further propose a method to utilize the proposed estimator to design the UHF for a wiretap code or secret key generation design. Simulation results show that prior methods significantly underestimate the M.I., particularly when using UHF for higher blocklengths (n ≫16). The proposed method can scale the blocklength up to 255, and we conjecture that the design can scale well to even higher blocklengths given adequate training data and model size. Additionally, we contend that our proposed estimator and adaptive hash design framework offer a practical approach for extending PLS considerations for wiretap channels into the finite blocklength regime.
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 0.98/0.76/0.050.98 / 0.76 / 0.050.98/0.76/0.05 ): 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)(Y_{E},H_{E})(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\tau = 0.27τ=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.
Physical-layer security (PLS) is a promising technique to complement more traditional means of communication security in beyond-5G wireless networks. However, studies of PLS are often based on ideal assumptions such as infinite coding blocklengths or perfect knowledge of the wiretap link’s channel state information (CSI). In this work, we study the performance of finite blocklength (FBL) transmissions using a new secrecy metric — the average information leakage (AIL). We evaluate the exact and approximate AIL with Gaussian signaling and arbitrary fading channels, assuming that the eavesdropper’s instantaneous CSI is unknown. We then conduct case studies that use artificial noise (AN) beamforming to analyze the AIL in both Rayleigh and Rician fading channels. The accuracy of the analytical expressions is verified through extensive simulations, and various insights regarding the impact of key system parameters on the AIL are obtained. Particularly, our results reveal that allowing a small level of AIL can potentially lead to significant reliability enhancements. To improve the system performance, we formulate and solve an average secrecy throughput (AST) optimization problem via both non-adaptive and adaptive design strategies. Our findings highlight the significance of blocklength design and AN power allocation, as well as the impact of their trade-off on the AST.
This letter considers a data-driven approach to a secret key rate analysis of physical layer secret key generation based on mutual information neural estimator (MINE), which achieves the state-of-the-art estimation accuracy. Since MINE does not require any statistical knowledge of randomness sources, the proposed approach is applicable to any source model and even real data which cannot be fully expressed by a simple model. We first demonstrate by simulations that the proposed estimation provides good approximations to the theoretical secret key rate for a simple model, and then show that the achievable secret key rate highly depends on sources of randomness and employed channel estimation schemes for a realistic wireless channel model. As an application of our analysis, we present a method to train quantization through backpropagation that maximizes the estimated secret key rate.
This paper presents a secrecy performance study of a wiretap communication system with finite blocklength (FBL) transmissions over Rayleigh fading channels, based on the definition of an average information leakage (AIL) metric. We evaluate the exact and closed-form approximate AIL performance, assuming that only statistical channel state information (CSI) of the eavesdropping link is available. Then, we reveal an inherent statistical relationship between the AIL metric in the FBL regime and the commonly-used secrecy outage probability in conventional infinite blocklength communications. Aiming to improve the secure communication performance of the considered system, we formulate a blocklength optimization problem and solve it via a low-complexity approach. Next, we present numerical results to verify our analytical findings and provide various important insights into the impacts of system parameters on the AIL. Specifically, our results indicate that i) compromising a small amount of AIL can lead to significant reliability improvements, and ii) the AIL experiences a secrecy floor in the high signal-to-noise ratio regime.
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.
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.