Training School: Enabling Multilingual Eye-Tracking Data Collection for Human and Machine Language Processing Research

With the support of the MultiplEYE COST Action CA21131, in Kaunas Technical University, organized a three-day training school in Lithuania, on April 27-28. Kaunas University of Technology (KTU) was founded in 1922. Currently, 7298 students are studying at KTU, of whom 1141 are international students. The main purpose of the training school was to introduce eye-tracking methodology, conduction of an eye-tracking experiments, demonstration of eye-tracking devices and software implementation of eye-tracking experiments. Ramunė KASPERĖ, leader of WG1 with her excellent team hosted the training school in the Faculty of Social Science, Arts and Humanities. 73 participants participated in the training school where11 training courses including hands-on experience were presented.

On the first day, the action vice chair Lena Jäger and WG1 leader and local host, Ramunė KASPERĖ welcomed the
training school participants in the spacious hall of the university. The first trainer Irina Sekerina, presented eyetracking methodology and provided several examples, papers and use cases. The next trainer Lena Jäger,
presented data collection, format of stored data and guidelines on conducting an eye-tracking experiment. Last
presentation of the first day was titled “Self-paced reading: practical introduction” by Jan Chromy who provided
pros and cons of self-paced reading experiments.

On the second day, Cengiz Acartürk introduced experimental design, providing cases using eye-tracking for
multimodal data. Moreover, the workshop by Pawel Kasprowski explained several types of eye trackers and
provided hands-on experience to play in eye-controlled games. Pawel drew attention to the need to comply with
the conditions of the experiment. Finally, the workshop coordinated by Deborah Jakobi, introduced MultiplEYE
experiment. She taught participants how to set up an eye-tracking experiment with Python using anaconda
environment. The training school participants followed the instructions and tested the working code for the
MultiplEYE experiment.

The last day started with introducing pymovements library by the trainer Daniel Krakowczyk in charge of software
development. Daniel conducted a step by step workshop: how to load the dataset, process eye tracking data,
shop plots and conducting research with pymovements. Finally, the training school presentations were closed
with Cengiz Acartürk with practical recommendations for setting up an eye-tracking lab.