Sorry, you need to enable JavaScript to visit this website.

Ensemble and Personalized Transformer Models for Subject Identification and Relapse Detection in E-Prevention Challenge

Citation Author(s):
Salvatore Calcagno, Raffaele Mineo, Daniela Giordano, Concetto Spampinato
Submitted by:
Salvatore Calcagno
Last updated:
25 May 2023 - 7:37am
Document Type:
Presentation Slides
Document Year:
Salvatore Calcagno
Paper Code:


In this short paper, we present the devised solutions for the subject identification and relapse detection tasks, which are part of the e-Prevention Challenge hosted at the ICASSP 2023 conference [1] [2] [3]. We specifically design an ensemble scheme of six models - five transformer-based ones and a CNN model - for the identification of subjects from wearable devices, while a personalized - one for each subject - scheme is used for relapse detection in psychotic disorder.
Our final submitted solutions yield top performance on both tracks of the challenge: we ranked 2nd on the subject identification task (with an accuracy of 93.85%) and 1st on the relapse detection task (with a ROC-AUC and PR-AUC of about 0.65).
Code and details are available at

0 users have voted:


Presentation Slides