A Resilient Collaborative Detection And Decision Framework Based On Ai To Enhance Security Against...

Universities and Institutes of France

France

October 21, 2022

Description

  • Organisation/Company: University of Bourgogne et ERICSSON FRance
  • Research Field: Computer science › Informatics Engineering › Communication engineering
  • Researcher Profile: First Stage Researcher (R1) Recognised Researcher (R2) Established Researcher (R3) Leading Researcher (R4)
  • Application Deadline: 21/10/2022 00:00 - Europe/Brussels
  • Location: France › Massy Plaiseau
  • Type Of Contract: Temporary
  • Job Status: Full-time
  • Short description:

    The main purpose of this PhD thesis is to propose and develop innovative collaborative detection (prediction) and decision-making techniques based on machine learning algorithms to protect the critical components of 5G's RAN from smart and complex attacks such as AI-related attackers and unknown threats. Among the main components of 5G's RAN that are attractive targets of attackers, we cite Control Unit (CU), Decision Unit (DU), Radio Unit (RU).

    The idea is that the AI detection and decision systems that will be proposed by the PhD will be activated at each critical virtual function and collaborate between each other to detect the unknown attacks' misbehavior (i.e., zero-day attacks), while taking into account the network metrics such as latency, communication overhead and packets lost. The expected results of the PhD can be summarized as follows:

  • Propose new AIrelated attacks models of the B5G's RAN.
  • Propose resilient collaborative hybrid detection systems able to detect the known and unknown attacks' misbehaviors and to be resilient against the AIrelated attackers targeting the critical components of 5G's RAN (where the detection system is activated).
  • Propose a mathematical model of collaborative cyber decisionmaking systems. This model investigates the behaviors of suspected attackers by monitoring the interaction between the hybrid detection system and these attackers with the goal to refine the detection provided by the hybrid system, i.e., reduces further the false positive rate.
  • Conceive a Proof of Concept (PoC) for the resilient collaborative detection (prediction) and cyber decisionmaking systems that take into account the security and B5G network metrics, such as detection and false positive rates, reaction time, latency, computation overhead and packets lost. The PoC will be embedded within Virtual Network Functions (VNFs) deployed within testbed network (such as Open-Air Interface).
  • Interact with 3GPP Ericsson experts (SA5 and SA3) for a possibility to standardize a part or all the software building blocks of the resilient collaborative detection and cyber decisionmaking systems.
  • The main innovative aspect of this PhD thesis is to study the optimal combination between the signature-based detection and machine learning based detection techniques with a goal to leverage the advantages of each detection technique against unknown threats and to be resilient from AI-related attacks. In addition, the PhD thesis will focus on proposing a new reaction mechanism based on a decision –making model (e.g., by using game theory) to address the decision-making issue and hence reduce further the false positive rate.

    Funding category: Cifre

    ERICSSON

    PHD title: Informatique

    PHD Country: France

    Offer Requirements Specific Requirements

    - MSc level in a technical field or an equivalent level of knowledge, especially in cyber security and AI.

    - Good understanding of Machine Learning theory and techniques

    - Good programming skills in Python, (R, Scala)

    - Applications/ domain-knowledge in telecommunication is a plus.

    - Well developed communication skills

    - Personal values in line with Ericsson core values

    - Large degree of flexibility and willingness to seek different tasks

    - Strong result oriented and finds it stimulating to work with change in a global team setup

    - Good English language skills in both writing and conversation, and additional French language skills are a plus

    Contact Information
  • Organisation/Company: University of Bourgogne et ERICSSON FRance
  • Organisation Type: Public Research Institution
  • Country: France
  • Similar Jobs

    Add to favorites Read more...

    Software Developer (M/F/D) In The Field Of Cryptography

    We are Austria's largest research and technology organisation and an Security is seeking to hire new Ingenious Partners for our location in Software Developer (m/f/d) in the field of Cryptography Cryptography is a key enabler...

    Add to favorites Read more...

    Interaction Specification Mining With Adversarial Examples

    A 3-year long full-time PhD student in Softtware Engineering on the following Interaction specification mining with adversarial examples component-intensive software systems, particularly those composed of execution logs of the communicating components. objective of the doctoral...

    Add to favorites Read more...

    Advanced Machine Learning

    Location: Italy › Bolzano Research in the scientific-disciplinary field covered by the call with Organisation/Company: Libera Università di Bolzano Organisation/Company: Libera Università di Bolzano

    Mälardalen University

    Sweden Sep 21, 2022

    Add to favorites Read more...

    Researcher in Probabilistic Real-Time Systems

    Researcher in Probabilistic Real-Time Systems Västeras Mälardalen University is growing and evolving due to increased research we can create a sustainable future through knowledge and innovation. all disciplinary domains and our research in Future Energy...