ANDREAS FISCHER

Two publications of our group have been accepted in this years ACIT conference in Budweis. One is on the generation of synthetic training data for process mining. The other deals with clustering approaches for word sense induction. See details in the publications list.

A new publication on Contextual Categorization Enhancement through LLMs Latent-Space has been presented in Venice at the IARIA COMPUTATION TOOLS 2024 conference. See details in the publications list.

Two new publications have been accepted recently. One conference paper to the ACIT conference in Wroclaw on "Analysis of Features-Brightness Bias in Computer Vision" and a journal article in the Bavarian Journal of Applied Sciences on "Topical Clustering of Unlabeled Transformer-encoded Researcher Activity". We're very glad to be successful once, again. See details in the publications list.

A new journal article on "Improving Image Tracing with Convolutional Autoencoders by High-Pass Filter Preprocessing" (a follow-up from a previous IARIA conference) is now published online. See details in the publications list.

I did some fixes to the website structure. There should be only few noticeable visible changes, but the underlying HTML and CSS are now defined better. In particular, I reduced the number of <h1> tags on pages, fixed some actual errors, and improved my knowledge on current HTML somewhat (I did not realize, ending with /> on a void element has actually become a discouraged practice).

Two of my papers have been accepted at the ICMECE Conference. The publications list is already updated, accordingly. Unfortunately, I won't make it to Barcelona myself, this time.

I have performed an extensive redesign of the site. Instead of using a Wordpress instance, the site now consists of static webpages generated by Pelican. This should improve performance and security. At the same time it gave me a chance to play around with a pet project.

The design is my own, including the new BibTeX renderer. Check out the list under Publications, which is dynamically generated from a BibTeX file.

Overall, I'm quite pleased with the result. A few bugs remain to iron out, but for now I'll call it a day.