From: 2024-01-18 11:15 to 12:00
Place: E:4130 (LUCAS)
Contact: birger [dot] swahn [at] cs [dot] lth [dot] se

One Computer Science MSc thesis to be presented on 18 January

Thursday, 18 January there will be a master thesis presentation in Computer Science at Lund University, Faculty of Engineering.

The presentation will take place in E:4130 (LUCAS).

Note to potential opponents: Register as an opponent to the presentation of your choice by sending an email to the examiner for that presentation (firstname.lastname@cs.lth.se. 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.)

11:15-12:00 in E:4130 (LUCAS)

Presenters: Felix Forsström, Maks Epsteins
Title: Finding Existing and Novel Errors in Heat Pumps Using Unsupervised ML
Examiner: Elin A. Topp
Supervisors: Jacek Malec (LTH), Vilhelm Åkerman (Robert Bosch AB)

This study explores the possibility for using unsupervised and semi-supervised machine learning models to find points corresponding to both known and novel errors within heat pump operational data. Four unsupervised and two semi-supervised models were tested and evaluated, which were performed on data processed using 4 distinct techniques. Findings revealed varying performance across models and configurations, lacking a singular model capable of consistently predicting known errors across all heat-pump systems studied. While certain configurations displayed promising performance for specific heat pumps, accurately predicting errors proved challenging without domain knowledge. Novel error detection was performed by clustering data points of similar characteristics, which may then be evaluated by a domain expert. Successful clustering of predicted errors was achieved for select systems, suggesting the potential viability of an unsupervised/semi-supervised anomaly detection approach with deeper domain knowledge focusing on a subset of the system.

Link to popular science summary: https://fileadmin.cs.lth.se/cs/Education/Examensarbete/Popsci/240118_11EpsteinsForsström.pdf