Information processing in cognitive systems & statistical methods
Summer term 2026
Guests are welcome! Feel free to stop by (even when it is only for a single talk).
Schedule
Currently we are finalizing the schedule for the current semester. Therefore, please stay tuned. If you expect to give a talk (e.g., because you are writing your BSc/MSc/PhD thesis in our group), please come to the planning session, such that we can schedule your talk. Deviations in day, time or location are marked in red.
- Fri 17.04.2026 14:15-15:45 A302, Sand 13
-
- Planning session
- Fri 24.04.2026 14:15-15:45 A302, Sand 13
-
- postponed
- Fri 08.05.2026 14:15-15:45 A302, Sand 13
-
- Sascha Meyen (Research talk): An information-theoretic measure of confidence calibration
- Kriti Bhatia (Journal club): Kopiske, K. K. and Domini, F. (2018). On the response function and range dependence of manual estimation. Experimental Brain Research, 236, 1309–1320. DOI
- Fri 15.05.2026 14:15-15:45 A302, Sand 13
-
- postponed
- Fri 22.05.2026 10:15-11:45 F122, Sand 6 (note different time & location)
-
- Şura Genç (University of Konstanz, Germany): Continuous assessment of flow state using subjective and physiological measures
- Fri 05.06.2026 14:15-15:45 A302, Sand 13
-
- postponed
- Fri 12.06.2026 14:15-15:45 A302, Sand 13
-
- Kriti Bhatia (Research talk): Does feedback improve accuracy in manual size estimation?
- Frieder Göppert (Journal club): Devezer, B., & Buzbas, E. O. (2026). The difference between ”replicable” and ”not replicable” is not itself scientifically replicable. https://arxiv.org/abs/2604.26268
- Fri 19.06.2026 13:30-17:00 Maria-von-Linden Str. 1 (note different time & location)
-
- PhD-day of the Department of Computer Science
- Fri 26.06.2026 14:15-15:45 A302, Sand 13
-
- Hamit Başgöl (Research talk): Detecting auditory and visual regularities within fast-paced sequences
- Sinan Kılıç (Planning talk for BSc thesis): Comparing methods of isolating phasic from tonic pupil responses
- Fri 03.07.2026 14:15-15:45 A302, Sand 13
-
- Linus Szillat (Research talk): On the relation of optional stopping and fixed-N designs
- Fri 10.07.2026 14:15-15:45 A302, Sand 13
-
- Charlott Wendt (Psychologische Hochschule Berlin): Exploring unconscious priming from a dual-tasking perspective.
- Florian Ebmeier (Research talk): Towards foundation models in heating systems.
- Fri 17.07.2026 14:15-15:45 Zoom: https://zoom.us/j/96614082576?pwd=VDczaHMrazI1T3gxM0poVGQ2blJLQT09
-
- Gary Fisk (Georgia Southwestern State University, Georgia, USA): Challenges for Demonstrating Unconscious Perception.
- Tue 21.07.2026 14:30-16:00 10-10/A12, Maria-von-Linden Str. 6 (note different time & location) ( NIP colloquium )
-
- Frieder Göppert (Research talk): Regions of decision
- Fri 24.07.2026 14:15-15:45 A302, Sand 13
-
- Alina Ahrens (University of Bremen): TabPFN: The knight in shining armor?
Abstract: The complexity of data sets that psychologists must analyse has increased drastically over the past decades. While this offers great opportunities for research, it often comes with the issue that most psychologists are not trained to perform more complex analyses, such as machine learning methods, and lack the computational resources for elaborate hyperparameter tuning. TabPFN is a pre-trained neural network whose developers claim it is capable of making regression and classification predictions as good as state-of-the-art machine learning approaches, while being easier to implement and not requiring hyperparameter tuning or training. In a benchmark study, we systematically compared TabPFN's performance to that of penalised regressions, random forests, and gradient boosting decision trees across 15 diverse psychological data sets to evaluate whether it can keep its promise. - Postponed:
Alexander Blöck (Alumnus talk): Steady-state visually evoked potential decoding
- Alina Ahrens (University of Bremen): TabPFN: The knight in shining armor?
- Send a PDF-file of an (almost) final version of your talk by email to V. Franz a day before the talk (latest: 2h before the talk). Details for the PDF-file: 1 slide per page. Make sure that you do NOT create separate pages for each step of animations. Give this PDF-file a sensible name. E.g., colloq-(your-last-name)-(date).pdf. If you made major changes to the talk after sending it to V. Franz, then please also send the final version after your talk.
- Practice your talk!
- Adhere to the time-limits during your talk. Practice that!
- Present data as graphs (supplemented but not supplanted by numerical statistics). Often these graphs will simply be means with error-bars showing the standard error of the mean.
- Provide your name, the date of your talk, your institution (often this is simply: University of Tübingen), etc. at the title-slide.
- Practice your talk!
- Finally (just in case, I forgot to mention): Practice your talk! Send a PDF before the talk!
- General idea: In the journal club a member of our group present an influential, scientific article relevant to our current work. Articles should typically be recent (e.g., 3-5 years), but could also be older if of special interest.
- Duration: 5-10 min presentation + 25-20 min discussion (in total app. 30 min; please make sure you adhere to these time-limits!).
- Obtaining articles: Articles will be available at our file-server (with the path being e.g., EC-STORE/literature/articles/journal-club-WS2025), please ask a member of our group if you do not know how to access those.
- Formatting details: Please make sure that a meaningful reference (containing title, author, year, journal) is presented at this web-page (either by you or by sending an email to V. Franz) and that the full APA-reference in the correct APA-formatting is present on the title-slide of your presentation (besides the typical things that should always be on a title slide: your name, the date of your talk, your institution (often this is simply: University of Tübingen).
