Large-Scale Pre-Training for Goal-Directed Dialog

SUPR uses JavaScript for certain functions. We cannot guarantee that you will be able to use the system with JavaScript disabled.

Dnr:

NAISS 2024/22-855

Type:

NAISS Small Compute

Principal Investigator:

Sverker Sikström

Affiliation:

Lunds universitet

Start Date:

2024-06-15

End Date:

2025-07-01

Primary Classification:

50102: Applied Psychology

Webpage:

https://www.psy.lu.se/forskningsprojekt-ai-baserade-sprakmodeller-diagnos-av-depression-och-angest

Allocation

Alvis at C3SE: 250 GPU-h/month
Mimer at C3SE: 250 GiB

Abstract

The project is about assessment of mental health based on dialog between a patient and AI-bot. We will use GODEL (Grounded Open Dialogue Language Model), a large pre-trained language model for dialog. This model will be fine tuned to available dialog data on assessment of mental health.