Theoretical and Practical Issues in Feature Fusion hos Københavns Universitet

Machine learning

I sektionen arbejdes der med at analysere de eksplosivt voksende datamængder i vores samfund og udvikle de metoder, som gør computere i stand til at lære fra data, f.eks. Kunstig intelligens træning af computere ud fra data.

Forskningsfeltet opdeles på DIKU i fire hovedområder

  • Machine Learning Theory
  • Medical Imaging
  • Information Retrieval
  • Machine Learning in Biology

 

Læs mere her: https://di.ku.dk/english/research/machine-learning/

 

 

Engelsk version:

 

The activities in the section range from research into the theoretical foundations of machine learning to applications within a broad set of domains, including remote sensing, information retrieval, medical image analysis and modelling of biological data.

 

The research section is divided into four areas:

 

  • Machine Learning Theory
  • Medical Imaging
  • Information Retrieval
  • Machine Learning in Biology

 

Read more here: https://di.ku.dk/english/research/machine-learning/

 

 

 

roject idea:
Problem: how should features from different input modalities be fused together for classification tasks? What are the theoretical and practical properties of different fusion methods? Is there one definitively best method?

 

Skills:
Students should have a good background in mathematics and be able to program neural network models using pytorch.

 

Supervisor:
Desmond Elliott, de@di.ku.dk

 

Husk at nævne, at du fandt dette opslag på KU Projekt & Job