Stochastic Encoding for Secure and Covert Communication

You are reading version 1 of this direction, frozen 2026-10-04. The latest version is version 2.

Version 1 of 2194 papers242 conceptsfrozen 2026-10-04

In traditional cryptography, security is typically handled at the higher layers of the network stack, treating the physical communication channel as a "dumb" pipe that simply moves bits. However, a growing body of research suggests that we can achieve robust security and even total invisibility by moving the defense into the encoding process itself.

This direction, often termed stochastic encoding, involves injecting deliberate randomness into how information is mapped to signals. By turning a deterministic communication process into a probabilistic one, researchers are finding ways to ensure that even if an adversary intercepts a signal, they lack the "side information" necessary to distinguish the message from ambient noise.

The Intuition of One-to-Many Mapping

The fundamental building block of stochastic encoding is a shift from one-to-one mapping to one-to-many mapping. In a standard system, a message $M$ always results in codeword $X$. In a stochastic system, a message $M$ is associated with a "bin" of many possible codewords $\{X_1, X_2, \dots, X_n\}$. When it’s time to transmit, the system randomly picks one codeword from that bin.

As Xiaojun Tang 0001 and Ruoheng Liu demonstrate in Multiple Access Channels with Generalized Feedback and Confidential Messages, this "stochastic codebook binning" injects ambiguity. To a legitimate receiver with the right "key" or channel conditions, the specific codeword can be resolved. But to an eavesdropper, the intercepted signal appears as a blur of dummy information or artificial noise, maintaining high "equivocation"—a mathematical measure of the attacker's uncertainty.

This concept is extended to modern semantic communications by Shuo Shao 0001 and Qianqian Yang 0002 in A Nearly Information Theoretically Secure Approach for Semantic Communications over Wiretap Channel. They use superposition coding to hide semantic data within a random "constellation cloud," effectively making the intercepted signal equivalent to Gaussian noise for anyone without the proper decoding parameters.

Exploiting Physical Degrees of Freedom

Beyond abstract codebooks, researchers apply stochasticity to the physical properties of the signal—phase, power, and spatial direction.

  • Phase Randomization: Amir K. Khandani and E. Bateni show in A Secure Key Sharing Algorithm Exploiting Phase Reciprocity in Wireless Channels that rotating symbols by a shared, uniformly distributed phase value provides unconditional secrecy. Because the resulting signal is statistically independent of the original data, the eavesdropper gains zero information.
  • Spatial Scrambling: In the context of radar and 6G, Kai Wu and J. Andrew Zhang utilize "stochastic sign inversion" and random beam selection. In Green Joint Communications and Sensing Employing Analog Multi-Beam Antenna Arrays, they demonstrate that by randomly selecting sensing beams, a system can scramble signals in directions where no legitimate user is present, preventing eavesdroppers from intercepting meaningful data while the system continues to scan its environment.

Achieving Covertness through Uncertainty

While "secrecy" protects the content of a message, "covertness" protects the very fact that a message is being sent. Stochastic encoding is the primary tool for achieving this "low probability of detection."

The core challenge is distinguishing a signal from the background noise of the universe. Xinchun Yu explains in Finite Blocklength Analysis of Gaussian Random Coding in AWGN Channels Under Covert Constraint that high-dimensional Gaussian random coding can "shape" codewords to stay within a thin shell of power, avoiding the outliers that would tip off an adversary.

Furthermore, Sen Qiao and Daming Cao found that Alice can significantly boost covertness by exploiting an adversary's uncertainty about signal phase angles. In Covert Communication Gains from Adversary's Uncertainty of Phase Angles, they show that randomizing phase selections across a codeword forces an observer to average over all possible distributions, making it nearly impossible to distinguish legitimate communication from random fluctuations.

Randomizing the "How" of Encoding

Finally, some researchers are applying stochasticity to the internal logic of encoders to prevent structural attacks or reverse-engineering.

In the realm of compression, Jarek Duda 0001 proposes injecting pseudorandomness into the state transitions of entropy coders. As described in ANS-based Compression and Encryption with 128-bit Security, this prevents attackers from using statistical fingerprints or Markov chain analysis to reverse-engineer secret encoding tables.

Similarly, Kuan Cheng and Xin Li 0006 explore Locally Decodable Codes with Randomized Encoding. By using hidden randomness to apply permutations and masks to codewords, they transform worst-case adversarial error patterns into average-case random noise. This prevents an adversary from strategically targeting specific bits to disrupt the decoding process, effectively "blunting" the impact of a targeted attack.

Why It Works: The Power of Asymmetric Uncertainty

The recurring theme across these 148 papers is the creation of asymmetric uncertainty. By using private randomness—whether it's a shared secret key, a random seed in a neural network like HiDDeN by Jiren Zhu, or the natural reciprocity of a wireless channel—the transmitter ensures that the legitimate receiver has a "clearer" view of the randomness than the adversary.

As Bikash Kumar Dey and Sidharth Jaggi argue in The benefit of a 1-bit jump-start, and the necessity of stochastic encoding, in jamming channels, without this private randomness, a clever adversary can anticipate and neutralize any deterministic defense. Stochastic encoding ensures that the adversary is always one step behind, forced to guess against a distribution they cannot fully observe.

Papers in this version

Showing 60 of 194 papers.

History & citationVersion 2 · 2026-10-04

Current · Permanent link

  1. Version 2 (latest) · 2026-10-04 · 234 papers
  2. Version 1 · 2026-10-04 · 194 papers
Cite this direction

Latest version (a living page):

@misc{lacuna_stochastic_encoding_60eb8e00,
  title = {stochastic encoding},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/stochastic-encoding/rdr_60eb8e0009a044568c018b9d1da1963d},
  note = {Latest version of a living research direction; cite a version for a fixed text. Accessed 2026-11-20},
}

This exact version (version 1, frozen text):

@misc{lacuna_stochastic_encoding_for_secure_and_cover_60eb8e00_v1,
  title = {Stochastic Encoding for Secure and Covert Communication},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/stochastic-encoding-for-secure-and-covert-communication/rdr_60eb8e0009a044568c018b9d1da1963d/v/1},
  note = {Version 1, frozen 2026-10-04. Accessed 2026-11-20},
}