Cryptographic Challenges through Conditional Probability Inference
Session
Computer Science and Communication Engineering
Description
The rapid evolution of cyberattacks has rendered classical deterministic cryptographic methods insufficient, exposing vulnerabilities in ciphers once considered secure. To address these limitations, this paper investigates conditional probability as a unifying paradigm for modern cryptanalysis. The proposed approach integrates statistical techniques (frequency analysis, chi-square testing, and n-gram modeling), probabilistic inference (Bayesian updating and Markov chains), heuristic optimization, and machine learning models for automated key prediction. Experimental evaluation demonstrates that while frequency analysis alone achieves 62% decryption accuracy, chi-square testing raises this to 70%, Bayesian–Markov inference achieves 85%, and a hybrid machine learning framework reaches 92%. Unlike previous studies that treat these methods separately, this work introduces a unified hybrid framework that bridges classical statistical reasoning with modern AI-driven approaches. The results confirm that probabilistic reasoning significantly enhances both accuracy and efficiency compared to standalone methods. This contribution has direct implications for securing real-world applications in communication networks, financial transactions, and post-quantum cryptographic systems.
Keywords:
Cryptography, Conditional Probability, Cryptanalysis, Statistical Methods, Bayesian Inference, Machine Learning, Post-Quantum Security
Proceedings Editor
Edmond Hajrizi
ISBN
978-9951-982-41-2
Location
UBT Kampus, Lipjan
Start Date
25-10-2025 9:00 AM
End Date
26-10-2025 6:00 PM
DOI
10.33107/ubt-ic.2025.83
Recommended Citation
Vrellaku, Vesë; Berisha, Diellza; Ramosaj, Ela; Demolli, Stinë; Morina, Rron; and Brovina, Redon, "Cryptographic Challenges through Conditional Probability Inference" (2025). UBT International Conference. 15.
https://knowledgecenter.ubt-uni.net/conference/2025UBTIC/CS/15
Cryptographic Challenges through Conditional Probability Inference
UBT Kampus, Lipjan
The rapid evolution of cyberattacks has rendered classical deterministic cryptographic methods insufficient, exposing vulnerabilities in ciphers once considered secure. To address these limitations, this paper investigates conditional probability as a unifying paradigm for modern cryptanalysis. The proposed approach integrates statistical techniques (frequency analysis, chi-square testing, and n-gram modeling), probabilistic inference (Bayesian updating and Markov chains), heuristic optimization, and machine learning models for automated key prediction. Experimental evaluation demonstrates that while frequency analysis alone achieves 62% decryption accuracy, chi-square testing raises this to 70%, Bayesian–Markov inference achieves 85%, and a hybrid machine learning framework reaches 92%. Unlike previous studies that treat these methods separately, this work introduces a unified hybrid framework that bridges classical statistical reasoning with modern AI-driven approaches. The results confirm that probabilistic reasoning significantly enhances both accuracy and efficiency compared to standalone methods. This contribution has direct implications for securing real-world applications in communication networks, financial transactions, and post-quantum cryptographic systems.