The problem Interviews are great. You can learn a lot about a topic by talking to experts. If you are working on a product, talking to users is common advice. If you are a researcher, qualitative surveys using interviews are great fun. Transcribing interviews, however, sucks Doing it yourself is tedious and is a painfully acquired skill. Outsourcing work to one of the various transcription APIs also is no perfect solution....
I ask students that want to write a thesis with me to prepare an exposé (or project outline) before we start. For the student, this ensures they have an idea of the work involved and a reasonable timeline to complete it. For me, it provides a short overview about what we agreed on and highlights important deadlines. At a minimum, I think an exposé should be a roughly two-page PDF and include:...
Our paper “Challenges to Open Collaborative Data Engineering” got accepted to the Hawaii International Conference on System Sciences 2023 (HICSS 56). Sadly, we could not present in person due to a flight cancellation but we provided a short video presentation. Abstract Open data is data that can be used, modified, and passed on, for free, similar to open-source software. Unlike open-source, however, there is little collaboration in open data engineering. We perform a systematic literature review of collaboration systems in open data, specifically for data engineering by users, taking place after data has been made available as open data....
At the start of July, I visited the Softwarecampus Kickoff Event. The Softwarecampus is a project targeting self-described “EntrepreNerds” and sponsoring research projects with an industry partner with up to 100k€. Having been to a few ice-breaker events with fellow nerds I anticipated an awkward day but I was pleasantly surprised. In the hope of being pleasantly surprised again in the future, I wanted to share some ways the Softwarecampus team made the event stand out to me....
Abstract Belief change research investigates how agents adapt their knowledge with potentially conflicting information. A common formalization is by epistemic states, abstract entities often represented by faithful preorders. Operators describe how epistemic states change with new evidence and are classified by which postulates they satisfy. Different approaches have been suggested for the problem of iterated belief change. Recent work introduces uniform revision that revises an agent’s beliefs based on one static total preorder, therefore lowering representational costs....
Note: An extended version of this paper has been published at the 7th Workshop on Formal and Cognitive Reasoning, you can find it here. Our paper “Certification of Iterated Belief Changes via Model Checking and its Implementation” got accepted to the 19th International Workshop on Non-Monotonic Reasoning (NMR-2021) at the 18th International Conference on Principles of Knowledge Representation and Reasoning (KR 2021). During the workshop I held a short presentation of the results....
Our paper “On Using Model Checking for the Certification of Iterated Belief Changes” got accepted to the 7th Workshop on Formal and Cognitive Reasoning (FCR-2021) at the 44th German Conference on Artificial Intelligence (KI-2021). During the workshop I held a short presentation of the results. You can find both the paper as well as the presentation slides here. Abstract The theory of iterated belief change investigates how epistemic states are changed according to new beliefs....
I created this synopsis of some ideas from the paper “How to Revise a Total Preorder” by Booth and Meyer during the seminar on “Representation and processing of uncertain knowledge with logic-based methods” at the University of Hagen. The slides are from my final presentation of the topic (and might be more or less useless on their own). Abstract Adapting one’s world view in the light of new information is a central skill of intelligent agents....
The popularity of machine learning, data science and related disciplines is exploding and with it the amount of courses, books, block posts etc you are exposed to. I recently finished the relatively old but highly rated course Machine Learning by Stanford University on Coursera and wanted to take the chance to offer my review and notes I took. The course Although the course is old enough to be referred to as “classic” by quite a few descriptions I have read it is timeless in the sense that most good introductions are....