Background

In Ericsson when delivering code, there are a lot of automatic tests run. If something goes wrong in those tests, we manually investigate the error and try to connect it to already known issues.

We would like to create a machine learning algorithm that does most of that work for us. Based on the existing data and already solved/closed tickets we believe that a machine learning solution can classify new errors without manual intervention in most cases.

Thesis Description

The following steps are envisioned as part of the thesis work:

  • Get an understanding of the nature of the current data available.
  • Investigate similar research papers previously performed.
  • Investigate and conclude which machine learning algorithm that would be most suitable for the available data.
  • Implement algorithm and investigate if the result is reliable enough to use.

The thesis will be concluded with a result presentation for the Ericsson team.

Qualifications

This project aims at students in computer science, computer engineering, machine learning or similar. Background in wireless communication is preferred.

Extent

1-2 students, 30hp each

Location

Ericsson AB Mjärdevi, Linköping

Preferred Starting Date

Spring 2024

Keywords

Mobile Telecommunication, Optimization, AI/ML

Contact Persons

Camilla Bodin                                                                                                                Johnny Blid

+46 724 66 67 56                                                                                                         +46 761 49 70 72

camilla.bodin@ericsson.com                                                                                       johnny.blid@ericsson.com