Dear all,

A postdoctoral position on Cybersecurity/IA is available starting January 2021 or later.
Initially, the position is for one year, but it can be extended for an additional year by mutual consent.

Applicants for the post doc position are invited to send a letter of interest and CV,
as well as at least one name of reference for recommendations, by email to:

- Albert Bifet: albert.bifet@telecom-paris.fr
- Mounira Msahli: mounira.msahli@telecom-paris.fr



Post-Doc: Fraud Detection

Keywords: Machine Learning  Scam detection  Security

1 General Context


Telephony is now merging several security challenges. Most telephony services
can be monetized that is why it can attracts scammers. Basically, all actors in
telephony ecosystem can be target of security attacks. In many cases, there are
no clear law about fraud and becomes difficult to address. Fraud loss was estimated
by operators as 12 billion dollars based on Cyber Telecom Crime Report 2019.
We believe that the real losses were more important and estimated to 3% and 10% of
total revenue.

This post-doc discusses all possible telephony frauds that can target billing
services of mobile network and to detect malicious behavior. We could mention,
for example, the Phishing scams or vishing. This fraud can also be divided
into several categories, like Wangiri scams[MS19], Voicemail scams and Blackmail/
Ransom or the International Revenue Sharing Fraud [Sah+17].
It is a question of detecting abnormal behavior in network. We can start by
analysing and modelling two frauds:

• "Smishing" Fraud: and Vishing are age-old scams, they remain prevalent
and effective; especially, with mobile devices and users. The techniques
employed by the scammers may vary, end-users are lured into disclosing
or revealing key and sensitive information on the Internet and also over
the phone [Eze09].

• "Wangiri" Fraud: is a Japanese word meaning ‘one ring and drop’. It
relies on one single ring method for a quick way to make money. Missed
calls from unknown callers entice subscribers to call back unknowingly


premium numbers where they are deceived to stay on the line for as long
as possible in order to inflate their bill[Mai19].


2 Objective

The idea of this research is to use Machine Learning to detect fraud in telephony
network. In literature, there are many studies using supervised algorithms
considering measurements such as true positive rate and accuracy at a
chosen threshold such as the number of correct predicted instances divided by
total number of instances. In fraud detection, the cost of false positive and false
negative error can differ from case to case, and can change over time.
In fraud detection, a false negative error is usually more costly than a false
positive error. The major issue of this work is to find the suitable machine
learning algorithms to detect frauds. The work in this post-doc contributes to
provide novel security architecture based on Machine Learning algorithms for
telephony network.


3 Background of the candidate

We are looking for a candidate with a PhD in Computer Sciences with very good
background in cybersecurity in mobile networks. A background in Machine
learning is essential. She or He must have a good knowledge of frauds and
attacks in such networks. Knowledge in performance evaluation, optimization,
and modeling will be greatly appreciated as well as programming and simulation
skills.

4 Contacts

- Mounira Msahli: mounira.msahli@telecom-paris.fr
- Albert Bifet: albert.bifet@telecom-paris.fr


5 References

[Eze09] Priscilla Mateko Amanor Ezer Osei Yeboah-Boateng. “Phishing, SMiShing
Vishing: An Assessment of Threats against Mobile Devices”. In:
Journal of Emerging Trends in Computing and Information Sciences.
2009.
[Mai19] George Sammour Mais Arafat Abdallah Qusef. “Detection of Wangiri
Telecommunication Fraud Using Ensemble Learning”. In: 2019
IEEE Jordan International Joint Conference on Electrical Engineering
and Information Technology (JEEIT). 2019.
[MS19] Abdallah Qusef Mais Arafat and George Sammour. “Detection of
Wangiri Telecommunication Fraud Using Ensemble Learning”. In:
2019 IEEE Jordan International Joint Conference on Electrical Engineering
and Information Technology (JEEIT). 2019.
[Sah+17] Merve Sahin et al. “SoK: Fraud in Telephony Networks”. In: 2017
IEEE European Symposium on Security and Privacy EuroSP. 2017.
[Pra+20] Sathvik Prasad et al. “Who’s Calling? Characterizing Robocalls through
Audio and Metadata Analysis”. In: 29th USENIX Security Sympo-
sium. https://www.usenix.org/conference/usenixsecurity20/
presentation/prasad. 2020




Kind Regards
Mounira MSAHLI