Introduction to ALJI
The Active Listening Journal Interaction ( ALJI) project aims to create a program designed to support the counseling of young students with depression by privately responding to their expressive writing or personal journal entries. ALJI may be used as a widely accessible entry point into clinical mental health support systems, providing critical guidance to those with suffering health.
To achieve this, a crisis detection model must be created and tested on a number of personal journal texts. This model is designed to align as closely as possible to your judgment of the personal journal text (see the Machine Learning section). You are given the ALJI Label Helper program to help label the dataset, which will be used in the training and testing of the final ALJI program.
The texts in the current dataset are publicly sourced from the Scholastic Awards website. The “Personal Essay and Memoir” category encourages a large amount of creativity and expressiveness from the authors, and is promising as a preliminary dataset. While this dataset may prove to be dissimilar to actual private journal entries, the numerous benefits of a passive (non-interactive) ALJI program can still outweigh the risks (see the Expressive Writing and Human-Computer Interaction sections of the project background page).
The ALJI project is open-source to ensure transparency, encourage collaboration, and to respect the goals of the project. You can see all components and code related to the project here: https://github.com/sublime09/ALJI