DIMA: Digital Mathematics


PhD position in Graph-based Learning

  • In this project we will develop new AI systems that allow bringing the reasoning closer to how a human learns and reasons about the world. More specifically, we will target the following key characteristics where humans perform better at than traditional AI systems: 
    - humans can cope much better with hierarchical information structure;
    - humans are able to learn much faster and better from limited data, often exploiting previous knowledge;
    - humans are able to perform much more complex reasoning and go beyond simple pattern matching.
    As graphs allow to efficiently represent relational data, we will tackle the above problems using a graph-based learning approach. The developed methodology will be applied to scanned document understanding in particular. 
  • We are looking for a Master in Mathematics, Computer Science or Engineering with expertise or interest in image processing and machine learning in particular. Coding skills are required.
  • Interested candidates should submit a cover letter detailing background and interest in this position, a full C.V. and information about grades during each Bachelor and Master year via email to Ann.Dooms@vub.be before the end of February 2020.