Creating and sharing knowledge for telecommunications

Project: Machine Learning based Profiling for Internet Security

Acronym: MaLPIS
Main Objective:
Security is one of the main concerns in today’s Internet, as large-scale attacks become more and more frequent. This situation will only get worse in the future, due to: (i) the increasing percentage of ciphered traffic, and (ii) the increasing number of, potentially insecure, attached devices and protocols, fostered by the Internet-of-Things.

Project MaLPIS (short for Machine Learning based Profiling for Internet Security) will develop advanced machine learning techniques targeted to the detection of Internet attacks and the security profiling of Internet entities (users/devices). The techniques span three important aspects - the methods, (ii) their benchmarking, in face of imperfect ground-truths, and (iii) their scalability, to cope with big data infrastructures - and will be tested in two testbeds related with the Internet-of-Things, provided by MEO from Portugal Telecom.
Reference: PTDC/EEI-TEL/32454/2017
Funding: FCT/PTDC
Start Date: 01-10-2018
End Date: 30-09-2021
Team: Rui Jorge Morais Tomaz Valadas, Eunice Isabel Ganhão Carrasquinha Trigueirão, José Marcelino Pousa, Periyadurai Karuppusamy, Paulo Jorge Salvador Serra Ferreira, Antonio Manuel Duarte Nogueira
Groups: Network Architectures and Protocols – Lx, Telecommunications and Networking – Av
Local Coordinator: Rui Jorge Morais Tomaz Valadas

Associated Publications
  • 2Papers in Journals
  • A. Garcia, M.R.O de Oliveira, R. Valadas, P. Salvador, A. Pacheco, Detection of Internet-wide traffic redirection attacks using machine learning techniques, IET Networks, Vol. 12, No. 4, pp. 179 - 195, July, 2023,
    | Abstract
    | BibTex
  • F.M. Macedo, R. Valadas, E. Carrasquinha, M.R.O de Oliveira, A. Pacheco, Feature selection using Decomposed Mutual Information Maximization, Neurocomputing, Vol. 513, No. 0, pp. 215 - 232, November, 2022,
    | Abstract
    | BibTex