Institut für Technik der Informationsverarbeitung

In-database support for machine learning

In-database support for machine learning

.
.
.
.

Circumstances

Currently we face more and more sensors which collect data in vehicles. Automated driving, predictive maintenance and an optimized range extension for electric vehicles are the trends that can be supported by Big Data management in a car or in the backend. To manage this, Mimer SQL Real-Time is a predictable, scalable and zero maintenance database management system with hard real-time guarantees. Mimer SQL Real-Time is intended for both critical real-time embedded systems as well as for applications that require high performance data management.

 

 

Goals and Tasks

  • Survey and categorise existing commercial products and research proposals, in different lifecycles of ML development.
  • Summarise the commonly used ML algorithms, how they are or need to be supported in databases, and the necessary interface pieces.
  • Implement a use case to compare the selected databases, and provide suggestions on effective ML support.
  • Make an implementation of ML support in Mimer

 

 

Constrainst and Requirements

  • Student of ETIT, Mechatronics or comparable
  • Knowledge about data bases in general
  • Machine Learnin background
  • At least one stay in Uppsala (near Stockholm)
  • A tight cooperation with Mimer in Uppsala is expected”