IEF 2020 is organized by the Centre for Entrepreneurship of the University of Cyprus in collaboration with PwC Cyprus. The 2020 Forum will bring together AI experts, researchers and professionals, decision-makers, entrepreneurs and game-changers to discuss the Challenge of Artificial Intelligence and address the most defying questions around the future of AI and its impact on society, economy and politics.

Recently our collaborators from University of Cyprus presented their AI research results “DriveGuard: A deep learning technique for countering cyber-attacks against autonomous vehicle camera sub-systems in the context of the H2020 CARAMEL project” as part of the activities of the 5th innovation & Entrepreneurship Forum (IEF2020): “The Challenge of AI”

Abstract

In the context of CARAMEL, we investigate the effect of camera sensor attacks for visual AI tasks in autonomous vehicles and present research results for a deep learning technique called DriveGuard, that act as a defense and mitigation mechanism for autonomous vehicle perception systems. This approach can provide protection against scenarios where an attacker can gain access to vital vehicle components by taking advantage of over-the-air updates in order to instantiate attacks that are not immediately detectable or perceivable, in contrast to just switching components off. DriveGuard utilizes the convolutional autoencoder family of deep learning networks to build an approach for efficient image reconstruction and anomaly detection. The approach will be integrated into an embedded anti-hacking device that will be capable for passive detection of attacks on an autonomous vehicle’s visual perception modules.

You can find more information about this topic at the following video:

IEF 2020 is organized by the Centre for Entrepreneurship of the University of Cyprus in collaboration with PwC Cyprus. The 2020 Forum will bring together AI experts, researchers and professionals, decision-makers, entrepreneurs and game-changers to discuss the Challenge of Artificial Intelligence and address the most defying questions around the future of AI and its impact on society, economy and politics.

Recently our collaborators from University of Cyprus presented “A modular approach to detect GPS location spoofing attack in autonomous vehicles” as part of the activities of the 5th innovation & Entrepreneurship Forum (IEF2020): “The Challenge of AI”

Autonomous vehicles can interact with their environment without human interaction through built-in sensors such as radar, rgb, lidar, GPS, ultrasonic that interact with the ECU to make automatic decisions. Therefore, enabling improvements in fuel efficiency, reduction of travel time, reduction of accidents, among others.

However, this dependence on sensor readings makes autonomous vehicles susceptible to attacks. An attacker could send false GPS signals to deceive the autonomous vehicles.

In this context, it’s being presented a cost-effective and modular solution for the AVs to timely and reliably detect and mitigate the GPS spoofing attack. Specifically, we develop an in-vehicle detection scheme that leverages multi-sensor fusion coupled with Bayesian filtering to produce in real-time an alternative GPS-free location stream of the AV that is used to check the integrity of the GPS location stream.

We invite you to watch the video presentation.

IEF2020 – Home (c4e.org.cy)

MOBIGEN is a South Korean company that started its business in 2000, after it opened its R&D Center. Since its establishment, they have been awarded many distinctions due to its outstanding performance. They have been focusing on Big Data & OSS(Operations Support System) Solution Provider for Operators in Major Infra Structure – Telecom, Electricity & Utility, mostly with a focus on Research and Development. The core concept of Mobigen solutions focuses on the Real-time Processing of Massive Data.

ETRI The Electronics and Telecommunications Research Institute is a Korean government-funded research institution. ETRI makes contribution to the nation’s economic and social development through research, development, and distribution of industrial core technologies in the field of Information, Communications, Electronics, Broadcasting, and Convergence technologies.

Its Role in CARAMEL

The CARAMEL project held its seventh plenary meeting in an online format with representatives of each of the consortium’s member companies, and on this occasion we had the participation of the new official partners.

In such meting each of the consortium entity gave a brief introduction to its former company, and a brief explanation about its role in the project.

it was an outstanding introductory meeting among all participants.

Did you know that some of the most common used social media channels are being affected by the creation of false accounts that aims to bias the opinion of specific large groups of people by publishing fake news.

Have you ever thought how might be possible to detect such accounts?

One of CARAMEL’s partners  has just released a very interesting article about Exploring Adversarial Attacks and Defences for Fake Twitter Account Detection.

We encourage to read this interesting just released paper.

• Papandreou, Andreas, Andreas Kloukiniotis, Aris Lalos, and Konstantinos Moustakas. “Deep Multi-Modal Data Analysis and Fusion for Robust Scene Understanding in CAVs”. Edited by IEEE MMSP 2021 (October 2021).

  AbstractBibTeXEndNote

  Deep learning (DL) tends to be the integral part of Autonomous Vehicles (AVs). Therefore the development of scene analysis modules that are robust to various vulnerabilities such as adversarial inputs or cyber-attacks is becoming an imperative need for the future AV perception systems. In this paper, we deal with this issue by exploring the recent progress in Artificial Intelligence (AI) and Machine Learning (ML) to provide holistic situational awareness and eliminate the effect of the previous attacks on the scene analysis modules. We propose novel multi-modal approaches against which achieve robustness to adversarial attacks, by appropriately modifying the analysis Neural networks and by utilizing late fusion methods. More specifically, we propose a holistic approach by adding new layers to a 2D segmentation DL model enhancing its robustness to adversarial noise. Then, a novel late fusion technique has been applied, by extracting direct features from the 3D space and project them into the 2D segmented space for identifying inconsistencies. Extensive evaluation studies using the KITTI odometry dataset provide promising performance results under various types of noise.

  @article{papandreou2021multimodal,
  abstract = {Deep learning (DL) tends to be the integral part of Autonomous Vehicles (AVs). Therefore the development of scene analysis modules that are robust to various vulnerabilities such as adversarial inputs or cyber-attacks is becoming an imperative need for the future AV perception systems. In this paper, we deal with this issue by exploring the recent progress in Artificial Intelligence (AI) and Machine Learning (ML) to provide holistic situational awareness and eliminate the effect of the previous attacks on the scene analysis modules. We propose novel multi-modal approaches against which achieve robustness to adversarial attacks, by appropriately modifying the analysis Neural networks and by utilizing late fusion methods. More specifically, we propose a holistic approach by adding new layers to a 2D segmentation DL model enhancing its robustness to adversarial noise. Then, a novel late fusion technique has been applied, by extracting direct features from the 3D space and project them into the 2D segmented space for identifying inconsistencies. Extensive evaluation studies using the KITTI odometry dataset provide promising performance results under various types of noise.},
  author = {Papandreou, Andreas and Kloukiniotis, Andreas and Lalos, Aris and Moustakas, Konstantinos},
  editor = {2021, IEEE MMSP},
  keywords = {cybersecurity},
  month = {October},
  title = {Deep multi-modal data analysis and fusion for robust scene understanding in CAVs},
  year = 2021
  }
  

  %0 Journal Article
  %1 papandreou2021multimodal
  %A Papandreou, Andreas
  %A Kloukiniotis, Andreas
  %A Lalos, Aris
  %A Moustakas, Konstantinos
  %D 2021
  %E 2021, IEEE MMSP
  %T Deep multi-modal data analysis and fusion for robust scene understanding in CAVs
  %X Deep learning (DL) tends to be the integral part of Autonomous Vehicles (AVs). Therefore the development of scene analysis modules that are robust to various vulnerabilities such as adversarial inputs or cyber-attacks is becoming an imperative need for the future AV perception systems. In this paper, we deal with this issue by exploring the recent progress in Artificial Intelligence (AI) and Machine Learning (ML) to provide holistic situational awareness and eliminate the effect of the previous attacks on the scene analysis modules. We propose novel multi-modal approaches against which achieve robustness to adversarial attacks, by appropriately modifying the analysis Neural networks and by utilizing late fusion methods. More specifically, we propose a holistic approach by adding new layers to a 2D segmentation DL model enhancing its robustness to adversarial noise. Then, a novel late fusion technique has been applied, by extracting direct features from the 3D space and project them into the 2D segmented space for identifying inconsistencies. Extensive evaluation studies using the KITTI odometry dataset provide promising performance results under various types of noise.
  

• Kantartopoulos, Panagiotis, Nikolaos Pitropakis, Alexios Mylonas, and Nicolas Kylilis. “Exploring Adversarial Attacks and Defences for Fake Twitter Account Detection”. Technologies 8, no. 4 (2020): 64. doi:https://doi.org/10.3390/technologies8040064.

  AbstractURLBibTeXEndNoteDOI

  Social media has become very popular and important in people’s lives, as personal ideas, beliefs and opinions are expressed and shared through them. Unfortunately, social networks, and specifically Twitter, suffer from massive existence and perpetual creation of fake users. Their goal is to deceive other users employing various methods, or even create a stream of fake news and opinions in order to influence an idea upon a specific subject, thus impairing the platform’s integrity. As such, machine learning techniques have been widely used in social networks to address this type of threat by automatically identifying fake accounts. Nonetheless, threat actors update their arsenal and launch a range of sophisticated attacks to undermine this detection procedure, either during the training or test phase, rendering machine learning algorithms vulnerable to adversarial attacks. Our work examines the propagation of adversarial attacks in machine learning based detection for fake Twitter accounts, which is based on AdaBoost. Moreover, we propose and evaluate the use of k-NN as a countermeasure to remedy the effects of the adversarial attacks that we have implemented.

  @article{kylilis2020exploring,
  abstract = {Social media has become very popular and important in people’s lives, as personal ideas, beliefs and opinions are expressed and shared through them. Unfortunately, social networks, and specifically Twitter, suffer from massive existence and perpetual creation of fake users. Their goal is to deceive other users employing various methods, or even create a stream of fake news and opinions in order to influence an idea upon a specific subject, thus impairing the platform’s integrity. As such, machine learning techniques have been widely used in social networks to address this type of threat by automatically identifying fake accounts. Nonetheless, threat actors update their arsenal and launch a range of sophisticated attacks to undermine this detection procedure, either during the training or test phase, rendering machine learning algorithms vulnerable to adversarial attacks. Our work examines the propagation of adversarial attacks in machine learning based detection for fake Twitter accounts, which is based on AdaBoost. Moreover, we propose and evaluate the use of k-NN as a countermeasure to remedy the effects of the adversarial attacks that we have implemented.},
  author = {Kantartopoulos, Panagiotis and Pitropakis, Nikolaos and Mylonas, Alexios and Kylilis, Nicolas},
  journal = {Technologies},
  keywords = {cybersecurity},
  number = 4,
  pages = 64,
  publisher = {Multidisciplinary Digital Publishing Institute},
  title = {Exploring Adversarial Attacks and Defences for Fake Twitter Account Detection},
  volume = 8,
  year = 2020
  }
  

  %0 Journal Article
  %1 kylilis2020exploring
  %A Kantartopoulos, Panagiotis
  %A Pitropakis, Nikolaos
  %A Mylonas, Alexios
  %A Kylilis, Nicolas
  %D 2020
  %I Multidisciplinary Digital Publishing Institute
  %J Technologies
  %N 4
  %P 64
  %R https://doi.org/10.3390/technologies8040064
  %T Exploring Adversarial Attacks and Defences for Fake Twitter Account Detection
  %U https://www.mdpi.com/2227-7080/8/4/64
  %V 8
  %X Social media has become very popular and important in people’s lives, as personal ideas, beliefs and opinions are expressed and shared through them. Unfortunately, social networks, and specifically Twitter, suffer from massive existence and perpetual creation of fake users. Their goal is to deceive other users employing various methods, or even create a stream of fake news and opinions in order to influence an idea upon a specific subject, thus impairing the platform’s integrity. As such, machine learning techniques have been widely used in social networks to address this type of threat by automatically identifying fake accounts. Nonetheless, threat actors update their arsenal and launch a range of sophisticated attacks to undermine this detection procedure, either during the training or test phase, rendering machine learning algorithms vulnerable to adversarial attacks. Our work examines the propagation of adversarial attacks in machine learning based detection for fake Twitter accounts, which is based on AdaBoost. Moreover, we propose and evaluate the use of k-NN as a countermeasure to remedy the effects of the adversarial attacks that we have implemented.
  

The CARAMEL project held its sixth plenary meeting in an online format with representatives of each of the consortium’s member companies.

This special session was attended by the external advisory board EAB, composed of international experts in the key areas for CARAMEL, who listened to the progress made in each of the work packages and provided strategic guidance and support, so that the project would achieve the proposed objectives.

On this occasion, the topics addressed were

• WP2 Use cases, risk assessment, requirements and architecture
• WP3 Countermeasures and mitigation techniques for advanced cybersecurity
• WP4 Cross-cutting detection and cybersecurity prevention
• WP5 Development of antihacking device and in-depth defense
• WP7 Dissemination, communication and exploitation of results