Towards a Latent Space Cartography of Subjective Experience in Mental Health.

Shawn Manuel ^1,2  Jean Gagnon ³  Frédéric Gosselin ³  Vincent Taschereau-Dumouchel ^1,2 

¹ Department of psychiatry and addictology, Université de Montréal   ² Centre de recherche de l’institut universitaire en santé mentale de Montréal   ³ Department of psychology, Université de Montréal  

Introduction

This repo, named DeepProjection, contains the code for our paper Towards a Latent Space Cartography of Subjective Experience in Mental Health. The latent space of subjective experiences can be explored on our companion website, as well as the synthetic stimulus generation (hosted on Hugging Face).

Directory Structure

DeepProjection/
├── assets/
│ ├── latentspace.png
│ └── trispace.png
├── src/
│ ├── clf.py
│ ├── fa.R
│ ├── gensynth.py
│ ├── plotting.py
│ └── scoring.py
├── LICENSE
├── README.md
└── requirements.txt

Detailed Descriptions

• assets/: Contains image files used for this README file.

  • latentspace.png: Project emblem
  • trispace.png: Project banner

• src/: Contains source code for the project.

  • clf.py: Script for valence and arousal classification, as well as mental health factors
  • fa.R: Script for factor analysis, including scree plot and factor loading plot
  • gensynth.py: Script for generating synthetic stimuli
  • plotting.py: Script for plotting results
  • scoring.py: Script for scoring questionnaires

• LICENSE: The license under which the project is distributed.

• README.md: The main README file you are currently reading.

• requirements.txt: Lists the Python packages required to run the project.