Meet Klas Holmberg

Klas Holmberg is a Master’s student working on a project with Martin Rosvall and Infobaleen on product recommendation systems

When did you start your Master’s project?

February 2022.

What led you to do your master’s in IceLab, with Martin Rosvall?

I was enrolled in a project course that Martin was the Course Coordinator for called ‘Design-Build-Test’ and at the end of that course Martin sent me a message over canvas to invite me in for a talk about the coming future. I was looking for a thesis to work with so it all kind of ended up there.

What are you researching?

I am researching the viability of Node embedding algorithms for use in product recommendation systems.
A sort of sub-field of Graph Neural Networks (that does not actually use any neural networks).

Why is this important or interesting?

A lot of machine learning is based on the premise of manual feature engineering for maximizing specific classification tasks. This can lead to a lot of work and a lot of missed opportunities especially when working with highly heterogeneous datasets. The node embedding algorithms that I am working with take a more generalist approach to build an output that can be utilized independent of the downstream prediction task. It removes the manual feature engineering task from the workload and devices its own measure of similarity.

What do you think the most interesting thing is that you have learned so far about your project, or a result from your work?

Because most real world data is relationally complex and heterogeneous these algorithms seem to have had a huge impact on all sorts of fields, everything ranging from protein-protein interactions & drug side effects within biomedicine to fake news detection. This is something I find very interesting, the applicability range of these algorithms.

Favorite part of working in IceLab?

The pitches that people produce for sure. Getting a first look into what people are researching and how it is progressing is really cool.

What are you thinking of doing next?

I actually just got an internship offer at Spotify for this summer as a Machine Learning Engineer. There I will be working with some cool Natural Language Processing tools and frameworks. It will be fun to get my hands on some real world data and see what I can do with that!

What do you nerd out on in your personal life?

One of my biggest hobbies is music; listening, writing, recording, rehearsing and playing live. I play in a band called Planetariet, a space themed Post-Rock instrumental band. Space is the place!
Besides that I never get tired of the philosophical conundrums of what Intelligence is to us and all the aspects of AI and what it will mean to our future as a society.

Excited about IceLab and wish you could join us as a team member? Our researchers often have opportunities for Master’s students, PhD scholarships and Postdoctoral Fellowships.  Browse our currently open positions or contact a researcher directly and see what options we have for you.

Share This Story, Choose Your Platform!