Call for Papers
First Workshop on Generalizability of Emotion Recognition from EEG Signals at ACII’24
The first workshop on Generalizability of Emotion Recognition from EEG Signals (EmoRec EEG) at ACII’24 is designed to improve the comparability and generalizability of EEG emotion recognition approaches. We invite submissions on emotion recognition from EEG signals in two tracks:
Challenge Track: In this track, we invite submissions addressing at least one of two well-defined EEG emotion recognition tasks. Submissions in this track may either have the form of a full paper (max. 7 pages main text + 1 page optional ethics statement + unlimited references) or of a short paper (max. 4 pages main text + 1 page references).
- Person-dependent task: In this task, participants will develop models that are trained and tested on the data of a specific participant.
- Person-independent task: Here, participants will develop approaches for EEG emotion recognition that generalize across users. A fixed train/test split will be used for evaluation.
For both tasks, we provide a comprehensive software framework for user-friendly pre-processing, training, and evaluation via the following link: https://github.com/EmotionLab/EEGain All models will be evaluated on four popular EEG emotion recognition datasets (DREAMER, MAHNOB-HCI, SEED, SEED IV). The framework supports dataset uploading in one line of code, but you need to have downloaded the datasets first. We have asked the owners of the datasets to promptly reply to future download requests.
Reflection Track: Here, we invite submissions concerned with the problem of generalizability of EEG emotion recognition. Both technical papers (max. 7 pages main text + 1 page optional ethics statement + unlimited references) and position papers (max. 2 pages main text + 1 page references) are possible. In contrast to the challenge track, the focus of technical papers is not to improve performance on the evaluation datasets, but to generate new insights into the problem. Evaluation on the challenge datasets is not required but will still be valued. Position papers may be concerned with all aspects pertaining to generalizability of EEG emotion recognition, including - but not limited to - evaluation protocols, data recording protocols, emotion elicitation techniques, EEG hardware considerations, and publishing bias.
Please note that all submitted papers should be anonymized for the review process.
Important Dates
Submission Deadline: June 12 2024 June 19 2024
Notification: July 14 2024
Workshop Day: September 15 2024
Submission
Papers can be submitted via ACII’s submission portal In the portal, you will have the option to specifically submit to the workshop.