Faculty of Engineering, LTH

Denna sida på svenska This page in English

Software@LTH events

CS MSc Thesis Presentations 16 June 2022


From: 2022-06-16 13:15 to 16:00
Place: E:4130 (Lucas)
Contact: birger [dot] swahn [at] cs [dot] lth [dot] se
Save event to your calendar

Two Computer Science MSc theses to be presented on 16 June

Thursday, 16 June there will be two master thesis presentations in Computer Science at Lund University, Faculty of Engineering.

The presentations will take place in E:4130 (Lucas). See time for each presentation below.

Note to potential opponents: Register as an opponent to the presentation of your choice by sending an email to the examiner for that presentation ( Do not forget to specify the presentation you register for! Note that the number of opponents may be limited (often to two), so you might be forced to choose another presentation if you register too late. Registrations are individual, just as the oppositions are! More instructions are found on this page.

13:15-14:00 in E:4130 (Lucas)

Presenter: Ebba Rickard
Title: Generating music using AI
Examiner: Elin Anna Topp
Supervisors: Emma Söderberg (LTH), Johan Davidsson (Axis Communications), Danny Smith (Axis Communications)

Network speakers are used in public spaces for public announcements and sometimes to play background music. To be able to provide alternative background music which does not require licensing, this thesis project explores the possibilities of generating music using machine learning methods. We trained and compared a set of generative models in the state-of-the-art. We then evaluated the results quantitatively and compared the outcome of the models based on this. 

Zoom link to presentation:

Link to popular science summary: To be updated


15:00-16:00 in E:4130 (Lucas)

Presenters: Fredrik Sjöström, Nicolas Petrisi
Title: ”First return, then explore" Adapted and Evaluated for Dynamic Environments
Examiner: Elin A. Topp
Supervisors: Volker Krueger (LTH), Hampus Åström (LTH)

Go-Explore is a new SOTA reinforcement learning, exploration, algorithm that beat the previous SOTA in 85.5% of the games in the Atari benchmark suite. However, due to the non-dynamic nature and lack of stochasticity, even with the presence of sticky actions and no-ops, Go-Explore lacks the ability to handle dynamic environments, such as random initial positions. Where a fixed, grid- based, cell representation is either misleading or inefficient and trajectories can not assume a fixed starting position; a dynamic form of representation is needed. Through alterations and additions to the algorithm, such as dynamic cells and OTF-trajectories, Go-Explore can be adapted to gain a more than seven times increased performance in an environment using random starting positions. Showing promising results for the future development of the algorithm.

Zoom link to presentation:

Link to popular science summary:öström.pdf