Event

ISE video channel on youtube: https://www.youtube.com/channel/UCjkkhNSNuXrJpMYZoeSBw6Q/

The ISE project lab is based on the summer semester lecture "Information Service Engineering". Goal of the course is to work on a given research problem in small groups (3-4 students) related to the ISE lecture topics, i.e. Natural Language Processing, Knowledge Graphs, and Machine Learning. The solution of the given research problem requires the development of a software implementation. 

The project will be worked on in teams of 3-4 students each, guided by a tutor from the teaching staff. 

Required coursework includes:

• Mid term presentation (5-10 min)
• Final presentation (10-15 min)
• Course report (c. 20 pages)
• Participation and contribution of the students during the course
• Software development and delivery

Notes:

The ISE project lab can also be credited as a seminar (if necessary).

The project will be worked on in teams of 3-4 students each, guided by a tutor from the teaching staff. 

Participation will be restricted to 16 students.

Participation in the lecture "Information Service Engineering" (summer semester) is required. There are video recordings on our youtube channel.

ISE Tutor Team:

• Dr. Genet Asefa Gesese
• M. Sc. Mirza Mohtasim Alam
• M. Sc. Oleksandra Bruns 
• M. Sc. Ebrahim Norouzi
• M. Sc. Mary Ann Tan
• B. Sc. Tabea Tietz
• M. Sc. Mahsa Vafaie 

WS 2023/24 Tasks List:

• Task 1: Zero-shot Ultrafine Typing of Named Entities. Use Pre-trained Language Models to assign predefined labels to entity mentions in a given context. Evaluate approaches which require no training data on a standard benchmark, i.e. UFET

• Task 2: Object Detection on Historical Theatre Photographs. Use Pre-trained DL models to detect and identify objects in historical theatre photographs and integrate the results into an existing Knowledge Graph.

• Task 3: Automatically Generate Ontologies from Competency Questions using Language Models. Competency questions (CQs) define the scope of knowledge represented in an ontology and are used to evaluate an ontology based on its ability to answer each question. In this task, we are investigating the benefit of Large Language Models to generate and evaluate ontologies from a set of competency questions.

• Task 4: Boosting the Performance of Large Language Models for Question Answering with Knowledge Graph Integration. Often, large language models hallucinate users with wrong or confusing answers. In order to generate relevant answers, knowledge graphs can help in many ways. The goal of this task is to utilize a knowledge graph to provide context and factual information to a language model, thereby improving the relevance and accuracy of its responses. 

• Task 5:Information Extraction and Knowledge Graph Engineering on the Use Case of Historical Political Flyers
  Information extraction and Knowledge Graph construction from digitized political leaflets of the Weimar Republic.

• Task 6: Sentiment Analysis on Multilingual Wikipedia. Analyse how different language Versions of Wikipedia differ in terms of Sentiment Bias.