English title
Deep learning in psychotherapy: Machine learning applied on therapeutic session recordings
Type
grant
Keywords

Psychotherapy process; Feedback-Informed Treatment; Routine Outcome Monitoring;
Routine Process Monitoring; Automatic Speech Recognition; Unsupervised Adaptation
of Speech Recognition system; Diarization; Natural Language Processing; Machine
Learning

Abstract

Psychotherapy is an expert activity requiring continuous decision-making and
continuous evaluation of the course of the psychotherapeutic process by the
psychotherapist. In practice, however, psychotherapists suffer from a lack of
immediate feedback to support this decision. The project aims to create a tool
that enables automated analysis of audio recordings of psychotherapeutic sessions
to provide psychotherapists feedback on the course in a short time. The project
is designed in cooperation with Brno University of Technology and Masaryk
University and is based on technologies of automatic speech recognition, natural
language computer processing, machine learning, expert coding of
psychotherapeutic process and self-assessment questionnaire methods. Its expected
outcome will be software providing psychotherapists with user-friendly and
practically beneficial feedback with the potential to improve psychotherapeutic
care.

Team members
Matějka Pavel, Ing., Ph.D. – research leader
Beneš Karel, Ing., Ph.D. (DCGM)
Burget Lukáš, doc. Ing., Ph.D. (DCGM)
Karafiát Martin, Ing., Ph.D. (DCGM)
Kašpárek Tomáš, Ing., Ph.D. (CVT)
Kesiraju Santosh, Ph.D. (DCGM)
Nehyba Jan, Mgr., Ph.D.
Novotný Ondřej, Ing., Ph.D.
Sarvaš Marek, Ing.
Žižka Josef, Ing. (DCGM)
Publications

2023

2021

Back to top