Thesis project 30hp - Online and Federated Learning

Background
Scania is one of the world’s leading manufacturer of trucks and buses for heavy transports, as well as industrial and marine engines. Transport services and logistics services make up an increasing part of our business, which guarantees Scania’s customers cost-efficient transport solutions and high availability. Over a million Scania vehicles are in active use, in over 100 countries.
In the Connectivity section within Scania R&D, we develop new solutions for connected vehicles in our Internet of Things (IoT) platform, as part of Scania’s increasing focus on communication, services, and smart transport solutions. Advanced data analysis capabilities are a cornerstone enabler in this development. 

Target/scope
Federated Learning is a promising method for training models for systems where data cannot be centrally stored, either due to privacy concerns and regulations or due to the technical/cost infeasibility of gathering it. This is the case for IoT devices coupled to non-stationary assets, such as heavy-duty vehicles. Training ML models using time series sensor data for predictive maintenance for heavy-duty vehicles in a federated learning fashion poses additional challenges, such as limited CPU, RAM and storage. One way to tackle these is to use online training. In this thesis you will combine state-of-the-art federated learning with online learning for anomaly detection.

1. Study literature on existing anomaly detection models, federated learning and online learning.
2. Build a bench setup with IoT devices that simulate streaming data
3. Setup the bench setup with Scaleout FEDn clients and FEDn studio
4. Benchmark and validate the federated learning bench the setup with open datasets
5. Implement online learning on the edge devices and benchmark the implementation against the previously build baseline
6. If time is available: use real truck data for anomaly detection

The student will be provided access to the computing infrastructure and to the dataset required for the task.

Education/line/direction
Area of education or direction: Masters programmes in Machine Learning, Data Science, Computational Mathematics, Complex Adaptive Systems, Computer Science or similar. 
Number of students: 1-2 
Start date for the Thesis project: January 2025
Estimated timescale: 20 weeks

Contact person and supervisor
Juan-Carlos Anderesen, Manager, 08-553 835 16, juan-carlos.anderesen@scania.com 

Application
Your application should contain the following: 
-  CV.
-  Personal letter. 
-  Copies of grades.
-  Optional: To propose tentative approach to the problem.
 
Date of publication
Until 2024-11-05. Applicants will be assessed on a continuous basis until the position is filled. Do not wait until the last date to apply. 
 

A background check might be conducted for this position. We are conducting interviews continuously and may close the recruitment earlier than the date specified.     

 

Thesis Worker