Pasar al contenido principal

Resultados relacionados

Resultados totales (Incluyendo duplicados): 12
Encontrada(s) 1 página(s)

UPCommons. Portal del coneixement obert de la UPC
Publicación Artículo científico (JournalArticle)

Low-cost training of image-to-image diffusion models with incremental learning and task/domain adaptation

  • Antona Pizà, Héctor
  • Otero Calviño, Beatriz|||0000-0002-9194-559X
  • Tous Liesa, Rubén|||0000-0002-1409-5843
Diffusion models specialized in image-to-image translation tasks, like inpainting and colorization, have outperformed the state of the art, yet their computational requirements are exceptionally demanding. This study analyzes different strategies to train image-to-image diffusion models in a low-resource setting. The studied strategies include incremental learning and task/domain transfer learning. First, a base model for human face inpainting is trained from scratch with an incremental learning strategy. The resulting model achieves an FID score almost equivalent to that of its batch learning equivalent while significantly reducing the training time. Second, the base model is fine-tuned to perform a different task, image colorization, and, in a different domain, landscape images. The resulting colorization models showcase exceptional performances with a minimal number of training epochs. We examine the impact of different configurations and provide insights into the ability of image-to-image diffusion models for transfer learning across tasks and domains., This work is partially supported by the Spanish Ministry of Science and Innovation under contracts PID2019-107255GB and PID2021-124463OB-IOO, by the Generalitat de Catalunya under grants 2017-SGR-962, 2021-SGR-00326, 2021-SGR-00478 and DRAC (IU16-011591). The research leading to these results has received funding from the European Union’s Horizon 2020 research and innovation program under the HORIZON-AG PHOENI2X (101070586) and HORIZON-EU VITAMIN-V (101093062) projects., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publicación Artículo científico (JournalArticle)

Automatic Linux malware detection using binary inspection and runtime opcode tracing

  • Alonso García, Martí
  • Gironès, Andreu
  • Costa Prats, Juan José|||0000-0003-2479-0230
  • Morancho Llena, Enrique|||0000-0003-2403-8145
  • Di Carlo, Stefano
  • Canal Corretger, Ramon|||0000-0003-4542-204X
The fast-paced evolution of cyberattacks to digital infrastructures requires new protection mechanisms to counterattack them. Malware attacks, a type of cyberattacks ranging from viruses and worms to ransomware and spyware, have been traditionally detected using signature-based methods. But with new versions of malware, this approach is not good enough, and new machine learning tools look promising. In this paper we present two methods to detect Linux malware using machine learning models: (1) a dynamic approach, that tracks the application executed instructions (opcodes) while they are being executed; and (2) a static approach, that inspects the binary application files before execution. We evaluate (1) five machine learning models (Support Vector Machine, k-Nearest Neighbor, Naive Bayes, Decision Tree and Random Forest) and (2) a deep neural network using a Long Short-Term Memory architecture with word embedding. We show the methodology, the initial dataset preparation, the infrastructure used to obtain the traces of executed instructions, and the evaluation of the results for the different models used. The obtained results show that the dynamic approach with a Random Forest classifier gets a 90% accuracy or higher, while the static approach obtains a 98% accuracy., We thank the anonymous reviewers for their thoughtful comments and suggestions which have improved the presentation of this work. This work is partially supported by the Spanish Ministry of Science and Innovation under contract PID2021-124463OB-IOO and by the Generalitat de Catalunya, Spain under grant 2021-SGR-00326. Finally, the research leading to these results has received funding from the European Union’s Horizon Programme under the HORIZON-EU VITAMIN-V (101093062) project., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publicación Artículo científico (JournalArticle)

A differential privacy protection-based federated deep learning framework to fog-embedded architectures

  • Gutiérrez Escobar, Norma
  • Otero Calviño, Beatriz|||0000-0002-9194-559X
  • Rodríguez Luna, Eva|||0000-0001-5904-7039
  • Utrera Iglesias, Gladys Miriam|||0000-0002-0637-1297
  • Mus León, Sergi
  • Canal Corretger, Ramon|||0000-0003-4542-204X
Nowadays, companies collect massive quantities of data to enhance their operations, often at the expense of sharing user sensible information. This data is widely used to train Deep Learning (DL) neural networks to model, classify, or recognize complex data. These activities enable companies to offer an array of services to users, such as precise advertising and optimal location services. This study explores potential solutions for preserving privacy while utilizing DL applications. To address the privacy issue, we develop a privacy-preserving framework specifically designed for fog computing environments. Unlike traditional cloud computing architectures, fog embedded architectures only share a small portion of user data with a nearby fog node, ensuring that the majority of sensitive data remains secure. Within these fog nodes, we incorporate two additional algorithms, namely Generalization and Threshold, to enhance the privacy-preserving capabilities of the framework. The first algorithm, Generalization, introduces a validation dataset within the fog nodes which not only increases the accuracy of the fog-embedded framework but also ensures that user data is preserved. The second algorithm, Threshold, is responsible for protecting user data samples and reducing the amount of information sent to the server. By combining these two algorithms, we are able to provide an additional layer of protection for user privacy while still maintaining the accuracy of the model. We conduct an evaluation to test its effectiveness using two separate datasets. In addition, we analyze them through a Feed Forward Neural Network (FFNN) and compare the results with a traditional centralized architecture to validate the effectiveness of the proposed framework. The results of our evaluation demonstrate that the proposed privacy-preserving framework, when combined with the Generalization and Threshold algorithms, can preserve up to 38.44% of user data. Additionally, we were able to extend the framework to multiple fog nodes without compromising the network’s accuracy, as we only observed a 0.1% decrease in accuracy when using the proposed architecture. This study emphasizes the importance of preserving user information while using DL applications and provides a solution that trains the desired network without violating user privacy, hence preserving their anonymity. Overall, the study highlights the potential of Federated Deep Learning to improve the accuracy and privacy of DL applications in fog computing environments., This work is partially supported by the Spanish Ministry of Science and Innovation, Spain under contracts PID2021-124463OB-IOO and PID2019-107255GB-C22; by the Generalitat de Catalunya, Spain under grant 2021SGR00326; and by the DRAC (IU16-011591), the HORIZON Vitamin-V, Spain(101093062), the HORIZON-AG PHOENI2X, Spain (101070586) and the HORIZON HORSE, Spain (101096342) projects., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publicación Artículo científico (JournalArticle)

Enhancing cybersecurity in railways: Machine learning approaches for attack detection

  • Otero Calviño, Beatriz|||0000-0002-9194-559X
  • Rodríguez Luna, Eva|||0000-0001-5904-7039
  • Costa Prats, Juan José|||0000-0003-2479-0230
  • Oriol, Mercedes
Ensuring the security of railway systems is crucial to protecting both passengers and infrastructure from the increasing threat of cyberattacks. As these threats grow in complexity and frequency, the need for resilient attack detection systems becomes more pressing. This study tackles the challenge of attack detection in railway systems using machine learning techniques. By integrating Generative Adversarial Networks (GANs), Convolutional Neural Networks (CNNs), and Transfer Learning (TL), we enhance detection accuracy and develop a robust approach capable of identifying both known and emerging threats. Our methodology utilized the Electra Modbus dataset as the source domain and was transferred to the Electra S7Comm dataset as the target domain. Experimental results highlight the effectiveness of GAN-based data augmentation in mitigating the scarcity of attack samples, enhancing both the robustness and generalizability of our detection models. Additionally, the application of pre-trained models through transfer learning played a crucial role in achieving superior performance. Our proposed solution can be deployed within railway networks to strengthen security measures, enabling proactive responses to cyber threats, safeguarding critical infrastructure, and ensuring uninterrupted operations., This work is partially supported by the Spanish Ministry of Science and Innovation under contract PID2021-124463OB-IOO and by the Generalitat de Catalunya, Spain under grant 2021-SGR-00326. Finally, the research leading to these results has received funding from the European Union’s Horizon 2020 research and innovation programme under the HORIZON-AG PHOENI2X (101070586), HORIZON-EU HORSE (101096342) and HORIZON-EU VITAMIN-V (101093062) projects., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publicación Artículo científico (JournalArticle)

Deep-learning based detection for cyber-attacks in IoT networks: A distributed attack detection framework

  • Jullian Parra, Olivia
  • Otero Calviño, Beatriz|||0000-0002-9194-559X
  • Rodríguez Luna, Eva|||0000-0001-5904-7039
  • Gutiérrez Escobar, Norma
  • Antona Pizà, Héctor
  • Canal Corretger, Ramon|||0000-0003-4542-204X
The widespread use of smart devices and the numerous security weaknesses of networks has dramatically increased the number of cyber-attacks in the internet of things (IoT). Detecting and classifying malicious traffic is key to ensure the security of those systems. This paper implements a distributed framework based on deep learning (DL) to prevent many different sources of vulnerability at once, all under the same protection system. Two different DL models are evaluated: feed forward neural network and long short-term memory. The models are evaluated with two different datasets (i.e.NSL-KDD and BoT-IoT) in terms of performance and identification of different kinds of attacks. The results demonstrate that the proposed distributed framework is effective in the detection of several types of cyber-attacks, achieving an accuracy up to 99.95% across the different setups., Open Access funding provided thanks to the CRUE-CSIC agreement with Springer Nature. This work is partially supported by the Spanish Ministry of Science and Innovation under contract PID2021-124463OB-IOO, by the Generalitat de Catalunya under grants 2017SGR962, 2021SGR00326, and by the DRAC (IU16-011591), the HORIZON Vitamin-V (101093062) and the HORIZON-AG PHOENI2X (101070586) projects., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publicación Artículo científico (JournalArticle)

A survey of machine and deep learning methods for privacy protection in the Internet of things

  • Rodríguez Luna, Eva|||0000-0001-5904-7039
  • Otero Calviño, Beatriz|||0000-0002-9194-559X
  • Canal Corretger, Ramon|||0000-0003-4542-204X
Recent advances in hardware and information technology have accelerated the proliferation of smart and interconnected devices facilitating the rapid development of the Internet of Things (IoT). IoT applications and services are widely adopted in environments such as smart cities, smart industry, autonomous vehicles, and eHealth. As such, IoT devices are ubiquitously connected, transferring sensitive and personal data without requiring human interaction. Consequently, it is crucial to preserve data privacy. This paper presents a comprehensive survey of recent Machine Learning (ML)- and Deep Learning (DL)-based solutions for privacy in IoT. First, we present an in depth analysis of current privacy threats and attacks. Then, for each ML architecture proposed, we present the implementations, details, and the published results. Finally, we identify the most effective solutions for the different threats and attacks., This work is partially supported by the Generalitat de Catalunya under grant 2017 SGR 962 and the HORIZON-GPHOENIX (101070586) and HORIZON-EUVITAMIN-V (101093062) projects., Peer Reviewed
Recercat. Dipósit de la Recerca de Catalunya
Publicación Artículo científico (JournalArticle)

Low-cost training of image-to-image diffusion models with incremental learning and task/domain adaptation

  • Antona Pizà, Héctor
  • Otero Calviño, Beatriz
  • Tous Liesa, Rubén
Diffusion models specialized in image-to-image translation tasks, like inpainting and colorization, have outperformed the state of the art, yet their computational requirements are exceptionally demanding. This study analyzes different strategies to train image-to-image diffusion models in a low-resource setting. The studied strategies include incremental learning and task/domain transfer learning. First, a base model for human face inpainting is trained from scratch with an incremental learning strategy. The resulting model achieves an FID score almost equivalent to that of its batch learning equivalent while significantly reducing the training time. Second, the base model is fine-tuned to perform a different task, image colorization, and, in a different domain, landscape images. The resulting colorization models showcase exceptional performances with a minimal number of training epochs. We examine the impact of different configurations and provide insights into the ability of image-to-image diffusion models for transfer learning across tasks and domains., This work is partially supported by the Spanish Ministry of Science and Innovation under contracts PID2019-107255GB and PID2021-124463OB-IOO, by the Generalitat de Catalunya under grants 2017-SGR-962, 2021-SGR-00326, 2021-SGR-00478 and DRAC (IU16-011591). The research leading to these results has received funding from the European Union’s Horizon 2020 research and innovation program under the HORIZON-AG PHOENI2X (101070586) and HORIZON-EU VITAMIN-V (101093062) projects., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Artículo científico (JournalArticle)

Automatic Linux malware detection using binary inspection and runtime opcode tracing

  • Alonso García, Martí
  • Gironès, Andreu
  • Costa Prats, Juan José
  • Morancho Llena, Enrique
  • Di Carlo, Stefano
  • Canal Corretger, Ramon
The fast-paced evolution of cyberattacks to digital infrastructures requires new protection mechanisms to counterattack them. Malware attacks, a type of cyberattacks ranging from viruses and worms to ransomware and spyware, have been traditionally detected using signature-based methods. But with new versions of malware, this approach is not good enough, and new machine learning tools look promising. In this paper we present two methods to detect Linux malware using machine learning models: (1) a dynamic approach, that tracks the application executed instructions (opcodes) while they are being executed; and (2) a static approach, that inspects the binary application files before execution. We evaluate (1) five machine learning models (Support Vector Machine, k-Nearest Neighbor, Naive Bayes, Decision Tree and Random Forest) and (2) a deep neural network using a Long Short-Term Memory architecture with word embedding. We show the methodology, the initial dataset preparation, the infrastructure used to obtain the traces of executed instructions, and the evaluation of the results for the different models used. The obtained results show that the dynamic approach with a Random Forest classifier gets a 90% accuracy or higher, while the static approach obtains a 98% accuracy., We thank the anonymous reviewers for their thoughtful comments and suggestions which have improved the presentation of this work. This work is partially supported by the Spanish Ministry of Science and Innovation under contract PID2021-124463OB-IOO and by the Generalitat de Catalunya, Spain under grant 2021-SGR-00326. Finally, the research leading to these results has received funding from the European Union’s Horizon Programme under the HORIZON-EU VITAMIN-V (101093062) project., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Artículo científico (JournalArticle)

Deep-learning based detection for cyber-attacks in IoT networks: A distributed attack detection framework

  • Jullian Parra, Olivia
  • Otero Calviño, Beatriz
  • Rodríguez Luna, Eva
  • Gutiérrez Escobar, Norma
  • Antona Pizà, Héctor
  • Canal Corretger, Ramon
The widespread use of smart devices and the numerous security weaknesses of networks has dramatically increased the number of cyber-attacks in the internet of things (IoT). Detecting and classifying malicious traffic is key to ensure the security of those systems. This paper implements a distributed framework based on deep learning (DL) to prevent many different sources of vulnerability at once, all under the same protection system. Two different DL models are evaluated: feed forward neural network and long short-term memory. The models are evaluated with two different datasets (i.e.NSL-KDD and BoT-IoT) in terms of performance and identification of different kinds of attacks. The results demonstrate that the proposed distributed framework is effective in the detection of several types of cyber-attacks, achieving an accuracy up to 99.95% across the different setups., Open Access funding provided thanks to the CRUE-CSIC agreement with Springer Nature. This work is partially supported by the Spanish Ministry of Science and Innovation under contract PID2021-124463OB-IOO, by the Generalitat de Catalunya under grants 2017SGR962, 2021SGR00326, and by the DRAC (IU16-011591), the HORIZON Vitamin-V (101093062) and the HORIZON-AG PHOENI2X (101070586) projects., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Artículo científico (JournalArticle)

A survey of machine and deep learning methods for privacy protection in the Internet of things

  • Rodríguez Luna, Eva
  • Otero Calviño, Beatriz
  • Canal Corretger, Ramon
Recent advances in hardware and information technology have accelerated the proliferation of smart and interconnected devices facilitating the rapid development of the Internet of Things (IoT). IoT applications and services are widely adopted in environments such as smart cities, smart industry, autonomous vehicles, and eHealth. As such, IoT devices are ubiquitously connected, transferring sensitive and personal data without requiring human interaction. Consequently, it is crucial to preserve data privacy. This paper presents a comprehensive survey of recent Machine Learning (ML)- and Deep Learning (DL)-based solutions for privacy in IoT. First, we present an in depth analysis of current privacy threats and attacks. Then, for each ML architecture proposed, we present the implementations, details, and the published results. Finally, we identify the most effective solutions for the different threats and attacks., This work is partially supported by the Generalitat de Catalunya under grant 2017 SGR 962 and the HORIZON-GPHOENIX (101070586) and HORIZON-EUVITAMIN-V (101093062) projects., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Artículo científico (JournalArticle)

Enhancing cybersecurity in railways: Machine learning approaches for attack detection

  • Otero Calviño, Beatriz
  • Rodríguez Luna, Eva
  • Costa Prats, Juan José
  • Oriol, Mercedes
Ensuring the security of railway systems is crucial to protecting both passengers and infrastructure from the increasing threat of cyberattacks. As these threats grow in complexity and frequency, the need for resilient attack detection systems becomes more pressing. This study tackles the challenge of attack detection in railway systems using machine learning techniques. By integrating Generative Adversarial Networks (GANs), Convolutional Neural Networks (CNNs), and Transfer Learning (TL), we enhance detection accuracy and develop a robust approach capable of identifying both known and emerging threats. Our methodology utilized the Electra Modbus dataset as the source domain and was transferred to the Electra S7Comm dataset as the target domain. Experimental results highlight the effectiveness of GAN-based data augmentation in mitigating the scarcity of attack samples, enhancing both the robustness and generalizability of our detection models. Additionally, the application of pre-trained models through transfer learning played a crucial role in achieving superior performance. Our proposed solution can be deployed within railway networks to strengthen security measures, enabling proactive responses to cyber threats, safeguarding critical infrastructure, and ensuring uninterrupted operations., This work is partially supported by the Spanish Ministry of Science and Innovation under contract PID2021-124463OB-IOO and by the Generalitat de Catalunya, Spain under grant 2021-SGR-00326. Finally, the research leading to these results has received funding from the European Union’s Horizon 2020 research and innovation programme under the HORIZON-AG PHOENI2X (101070586), HORIZON-EU HORSE (101096342) and HORIZON-EU VITAMIN-V (101093062) projects., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Artículo científico (JournalArticle)

A differential privacy protection-based federated deep learning framework to fog-embedded architectures

  • Gutiérrez Escobar, Norma
  • Otero Calviño, Beatriz
  • Rodríguez Luna, Eva
  • Utrera Iglesias, Gladys Miriam
  • Mus León, Sergi
  • Canal Corretger, Ramon
Nowadays, companies collect massive quantities of data to enhance their operations, often at the expense of sharing user sensible information. This data is widely used to train Deep Learning (DL) neural networks to model, classify, or recognize complex data. These activities enable companies to offer an array of services to users, such as precise advertising and optimal location services. This study explores potential solutions for preserving privacy while utilizing DL applications. To address the privacy issue, we develop a privacy-preserving framework specifically designed for fog computing environments. Unlike traditional cloud computing architectures, fog embedded architectures only share a small portion of user data with a nearby fog node, ensuring that the majority of sensitive data remains secure. Within these fog nodes, we incorporate two additional algorithms, namely Generalization and Threshold, to enhance the privacy-preserving capabilities of the framework. The first algorithm, Generalization, introduces a validation dataset within the fog nodes which not only increases the accuracy of the fog-embedded framework but also ensures that user data is preserved. The second algorithm, Threshold, is responsible for protecting user data samples and reducing the amount of information sent to the server. By combining these two algorithms, we are able to provide an additional layer of protection for user privacy while still maintaining the accuracy of the model. We conduct an evaluation to test its effectiveness using two separate datasets. In addition, we analyze them through a Feed Forward Neural Network (FFNN) and compare the results with a traditional centralized architecture to validate the effectiveness of the proposed framework. The results of our evaluation demonstrate that the proposed privacy-preserving framework, when combined with the Generalization and Threshold algorithms, can preserve up to 38.44% of user data. Additionally, we were able to extend the framework to multiple fog nodes without compromising the network’s accuracy, as we only observed a 0.1% decrease in accuracy when using the proposed architecture. This study emphasizes the importance of preserving user information while using DL applications and provides a solution that trains the desired network without violating user privacy, hence preserving their anonymity. Overall, the study highlights the potential of Federated Deep Learning to improve the accuracy and privacy of DL applications in fog computing environments., This work is partially supported by the Spanish Ministry of Science and Innovation, Spain under contracts PID2021-124463OB-IOO and PID2019-107255GB-C22; by the Generalitat de Catalunya, Spain under grant 2021SGR00326; and by the DRAC (IU16-011591), the HORIZON Vitamin-V, Spain(101093062), the HORIZON-AG PHOENI2X, Spain (101070586) and the HORIZON HORSE, Spain (101096342) projects., Peer Reviewed, Postprint (published version)