room 012 Roman Kern is the head of Knowledge Discovery at the Know-Center (competence centre for Big Data analytics and data-driven business) and works at the Institute for Interactive Systems and Data Science at the Technical University of Graz. He was awarded his Ph.D. by the Graz University of Technology. Before working in research he gained experience in industry projects as project manager, software architect and software engineer ranging from big and medium sized companies to small start-ups.
Roman's research interest are multi-disciplinary and include Natural Language Processing, Machine Learning and Information Retrieval - with a focus on Data Science and Big Data Analytics. He applies these methods in fields like Scientific Publication Mining, Intelligent Transportation Systems, and Smart Production. His work includes writing of proposals for national and international research projects; he served as coordinator, work package lead, scientist in charge and national contact point for numerous research projects, ranging from small national projects to big European projects.
In his work at the Know-Center his mission is to close the gap between science and industry via applied research projects, consulting and knowledge transfer. The demand in skills from the real world application scenarios also influences his teaching activities and shape the Computer Science curriculum of the Graz University of Technology.
List of courses at the Graz University of Technology.
There are three options available:
- Choose a topic, or propose an own topic and work on your own (see list below for options)
- Collaboration with local start-ups
- Work together with a (industrial, research) partner organisation.
Feel free to use the Latex thesis template (based on input from Karl Voit and Keith Andrews):
Collection of a few helpful tips for Master's thesis, provided by Annemarie Harzl. For printing the thesis one can choose the CopyShop or an online service (e.g., masterprint). After the thesis is finished beware of predatory publishers that offer to print the thesis for free!
List of open topics together with their domain - each topic can be chosen for either a Bachelor's or Master's thesis (emphasis and scope will be adapted)
- Recursive Word Sense Induction#nlp, #research
Collect a textual dataset, split and pre-process the data. Each word is then clustered, and pure clustered are used to split the word and replace its occurrence with a cluster representation. Continue this process until no pure clusters can be found.
- Learn to Stem#nlp, #research
Goal: Find out, which words should be stemmed, and which should remain intact in a preprocessing pipeline. Approach: Define a set of criteria (hypothesis) for making such a decision. Then, devise a evaluation design to find out, what the best criteria (or combination) is.
- Wikipedia Article Generator#nlp
Given just the infobox of an Wikipedia article, generate a full Wikipedia article.
- Authorship attribution based on style information#text-mining, #nlp
Can one detect the author just on the writing of a text?
- Speed-up symbolic regression with deep learning methods#deep-learning
e.g., use a deep learning model to restrict the search space of evolutionary algorithms (looks like sine-wave)
- Deep clustering for partitioned error propagation#deep-learning, #machine-learning
Can we improve the back-propagation by implicit clustering of activations/gradients?
- Classification of Wikipedia Pages#machine-learning, #data-science
e.g., identify pages about emergent technologies not labelled as such
Get an overview of the slots Have a look at: ISDS-KD-Seminar or as .ics. There you see when the next slots are and their availability (e.g., "2 slots free" in the title of the event).
Book a slot Just send me an e-mail with your preferred date, optimally including a title and a short abstract, I will then update the calendar.
Date/time/location/speakers Should be in the details of the event - occasionally, there will also be friendly reminders of upcoming presentations.
Give a presentation The presentation typically are 20 minutes, allowing for approx. 5 minutes of Q&A - the language by default is English (German only in exceptional cases). Please bring your own laptop for the presentation.
Please have a look at my Mendeley page or my Google Scholar page for a full list of publications.
List of scientific workshops.