Topological and Entropic Analysis of Steganography and Steganalysis via AI-Driven Multi-Agent Systems
Bibliographic record
Abstract
In this paper we introduce a novel combination of Topological data analysis (TDA) and the informationtheoretic entropy measures integrated into an AI-based Multi-Agent System (MAS) framework to provide an enhanced steganography and steganalysis analysis. As a consequence of advanced hidden means of the communication and the increased requirements for smart adaptive destruction methods, the convergence of topology, entropy and AI in security is studied. Leveraging core concepts from steganography, steganalysis, and MAS design, we introduce a model formed by special agents—Embedder, Extractor, Detector, and Coordinator—that circulate in centralized, decentralized, and hybrid communication models. These agents are augmented with deep learning models (CNNs, GANs, Trans- formers) to extract features, predict payloads, and detect anomalies. We use persistent homology to extract topological features from high dimensional stego-object Feature Spaces, manifesting difference in the underlying manifolds of cover vs. stego data. In addition, we use entropic measures (Shannon, Rényi, KL divergence, predictive entropy) of security, agent uncertainty and information loss/gain in communications. Synergy of topological and entropic views enable us to define joint topological-entropic signatures that enhance readability and the steganalysis agents' performance. Experimental on image and audio steganography data sets (e.g., BOSSBase, ALASKA) demonstrates the proposed method's superiority of detection accuracy and robustness over baseline methods. Our findings convey that MAS dynamics particularly in adversarial training, largely decide the topological complexity and entropy landscape of learned feature space. The paper finishes with some discussion on these results when designing adaptive AI-based steganalysis systems.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.003 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.000 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.001 | 0.002 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".