BEGIN:VCALENDAR
PRODID:-//eluceo/ical//2.0/EN
VERSION:2.0
CALSCALE:GREGORIAN
BEGIN:VEVENT
UID:36ff93a72300a1cb3badf135ee061bbc
DTSTAMP:20261013T001744Z
SUMMARY:CS MSc Thesis Presentation 17 February 2026
DESCRIPTION:Kontakt: birger.swahn@cs.lth.se\n\nTuesday\, 17 February there 
 will be a master thesis presentation in Computer Science at Lund Universit
 y\, Faculty of Engineering.The presentation will take place in E:4130 (Luc
 as).Note to potential opponents: Register as an opponent to the presentati
 on of your choice by sending an email to the examiner for that presentatio
 n (firstname.lastname@cs.lth.se). Do not forget to specify the presentatio
 n you register for! Note that the number of opponents may be limited (ofte
 n to two)\, so you might be forced to choose another presentation if you r
 egister too late. Registrations are individual\, just as the oppositions a
 re! More instructions for opponents are found here on the LTH thesis proje
 ct page.11:15-12:00 in E:4130 (Lucas)Presenters: Ebba Rakuljic Feldt\, Eli
 n HellströmTitle: Knowledge Graph–enhanced RAG for Enterprise Question 
 Answering systemsExaminer: Jacek MalecSupervisors: Marcus Klang (LTH)\, Na
 than Hoche (Axis Communications AB)As new knowledge emerges rapidly\, the 
 knowledge base of large language models (LLMs) can quickly become outdated
 \, requiring frequent retraining\, an expensive and resource-intensive pro
 cess. Retrieval-augmented generation (RAG) addresses this by providing LLM
 s with up-to-date or domain-specific knowledge at inference time. However\
 , traditional RAG systems struggle with broad\, summarizing queries and co
 mplex reasoning\, since retrieval is primarily based on semantic similarit
 y. Recently\, an alternative approach has been proposed that combines RAG 
 with a knowledge graph (GraphRAG) to improve logical consistency and reaso
 ning. In this thesis\, we compare the performance of GraphRAG and traditio
 nal RAG on both internal enterprise data from Axis and the GraphRAG-Bench 
 benchmark. For automatic construction of the knowledge graph\, we use LLMs
  to extract entities and relations. To evaluate the enterprise data\, we g
 enerate a synthetic question–answer dataset using LLMs and use an LLM-as
 -a-judge to assess performance. Our results show that\, under the evaluate
 d conditions\, traditional RAG achieves higher performance than GraphRAG o
 n many QA tasks\, both on the benchmark and in the LLM-based evaluation of
  internal data.Link to popular science summary\n\nMer information om händ
 elsen: https://www.cs.lth.se/evenemang/cs-msc-thesis-presentation-17-febru
 ary-2026
DTSTART;TZID=GMT:20260217T101500
DTEND;TZID=GMT:20260217T110000
LOCATION:E:4130 (Lucas)
END:VEVENT
END:VCALENDAR