Publications - Pervasive Computing Laboratory

Access control anotacionsemantica authentication Cloud computing coap compromise cryptography cynamon emadrid emrisco home network I-Shaper identity management inrisco inteligenciafuentesabiertas Internet of Things intrusion detection system IoT magos middleware Pervasive computing privacy Protocols Qursa raudo2 Security Smart Card Trust management Ubisec VANETs 
Show all
2025
 Blanco-Romero, Javier;  Lorenzo, Vicente;  Almenares-Mendoza, Florina;  Díaz-Sánchez, Daniel
Machine Learning Predictors for Min-Entropy Estimation Journal Article 
In: Entropy 2025, vol. 27, iss. 2, no. 156, pp. 1-31, 2025.
Abstract | Links | BibTeX | Tags: autoregressive processes, compromise, I-Shaper, machine learning predictors, min-entropy estimation, Qursa, random number generators
@article{javierblanco004,

title = {Machine Learning Predictors for Min-Entropy Estimation},

author = {Javier Blanco-Romero and Vicente Lorenzo and Florina Almenares-Mendoza and Daniel Díaz-Sánchez},

url = {https://www.mdpi.com/1099-4300/27/2/156},

doi = {https://doi.org/10.3390/e27020156},

year  = {2025},

date = {2025-02-02},

urldate = {2025-02-02},

journal = {Entropy 2025},

volume = {27},

number = {156},

issue = {2},

pages = {1-31},

abstract = {This study investigates the application of machine learning predictors for the estimation of min-entropy in random number generators (RNGs), a key component in cryptographic applications where accurate entropy assessment is essential for cybersecurity. Our research indicates that these predictors, and indeed any predictor that leverages sequence correlations, primarily estimate average min-entropy, a metric not extensively studied in this context. We explore the relationship between average min-entropy and the traditional min-entropy, focusing on their dependence on the number of target bits being predicted. Using data from generalized binary autoregressive models, a subset of Markov processes, we demonstrate that machine learning models (including a hybrid of convolutional and recurrent long short-term memory layers and the transformer-based GPT-2 model) outperform traditional NIST SP 800-90B predictors in certain scenarios. Our findings underscore the importance of considering the number of target bits in min-entropy assessment for RNGs and highlight the potential of machine learning approaches in enhancing entropy estimation techniques for improved cryptographic security.},

keywords = {autoregressive processes, compromise, I-Shaper, machine learning predictors, min-entropy estimation, Qursa, random number generators},

pubstate = {published},

tppubtype = {article}

}

Close
This study investigates the application of machine learning predictors for the estimation of min-entropy in random number generators (RNGs), a key component in cryptographic applications where accurate entropy assessment is essential for cybersecurity. Our research indicates that these predictors, and indeed any predictor that leverages sequence correlations, primarily estimate average min-entropy, a metric not extensively studied in this context. We explore the relationship between average min-entropy and the traditional min-entropy, focusing on their dependence on the number of target bits being predicted. Using data from generalized binary autoregressive models, a subset of Markov processes, we demonstrate that machine learning models (including a hybrid of convolutional and recurrent long short-term memory layers and the transformer-based GPT-2 model) outperform traditional NIST SP 800-90B predictors in certain scenarios. Our findings underscore the importance of considering the number of target bits in min-entropy assessment for RNGs and highlight the potential of machine learning approaches in enhancing entropy estimation techniques for improved cryptographic security.
Close
https://www.mdpi.com/1099-4300/27/2/156
doi:https://doi.org/10.3390/e27020156
Close