The Multimodal Sentiment Analysis Challenge (MuSe) 2023 is dedicated to multimodal sentiment and emotion recognition in different scenarios: https://www.muse-challenge.org

This year's edition features three different sub-challenges:


== Sub-Challenges, Datasets & Features ==

1. Emotional Mimicry Sub-challenge (MuSe-Mimic)
Predicting 3 discrete emotions (Approval, Uncertainty, Disappointment) from videos of mimicked emotional expressions. This sub-challenge utilises the novel Hume-Mimic dataset provided by Hume AI (https://www.hume.ai). It contains over 32 hours of audiovisual recordings of 557 different subjects, including transcripts.

2. Cross-Lingual Humor Detection Sub-challenge (MuSe-Humor)
Predicting the presence of humour in audio-visual recordings of football press conferences. In this sub-challenge, a cross-lingual setting is provided, where the training data comprises German recordings, while the test data is composed of English press conferences. Overall, more than 17 hours of data (audio, video, transcripts) are provided.

3. Personalisation Sub-challenge (MuSe-Personalisation)
Predicting emotional valence and psycho-physiological arousal signals from audio-visual recordings. In order to facilitate personalisation, parts of each test subject's labels are made available. MuSe-Personalisation uses the Ulm-TSST data set that includes audio, video, textual transcriptions and physiological signals (respiratory rate, ECG, BPM).

For the baseline paper see here: https://www.researchgate.net/publication/370100318_The_MuSe_2023_Multimodal_Sentiment_Analysis_Challenge_Mimicked_Emotions_Cross-Cultural_Humour_and_Personalisation

== How to Participate ==
Instructions are available at https://www.muse-challenge.org/challenge/participate. Data and features are available upon registration. Links to the baseline code and the preliminary baseline paper are also provided on the homepage (https://www.muse-challenge.org).


== Organisers ==
Shahin Amiriparian (University of Augsburg, GER)
Lukas Christ (University of Augsburg, GER)
Andreas König (University of Passau, GER)
Alan Cowen (Hume AI, USA)
Eva-Maria Meßner (University of Ulm, GER)
Erik Cambria (Nanyang Technological University/ SenticNet, SG)
Björn W. Schuller (Imperial College London/ audEERING, UK)