Modeling speech transcriptions for automatic assessment of depression severity
| buir.advisor | Dibeklioğlu, Hamdi | |
| dc.contributor.author | Kaynak, Ergün Batuhan | |
| dc.date.accessioned | 2022-09-20T11:14:22Z | |
| dc.date.available | 2022-09-20T11:14:22Z | |
| dc.date.copyright | 2022-09 | |
| dc.date.issued | 2022-09 | |
| dc.date.submitted | 2022-09-19 | |
| dc.department | Department of Computer Engineering | en_US |
| dc.description | Cataloged from PDF version of article. | en_US |
| dc.description | Thesis (Master's): Bilkent University, Department of Computer Engineering, İhsan Doğramacı Bilkent University, 2022. | en_US |
| dc.description | Includes bibliographical references (leaves 55-66). | en_US |
| dc.description.abstract | It is true that everyone has bad days from time to time. Unfortunately, for peo-ple suffering from depression, every day is a constant battle for motivation to do even the simplest of things, all the while dealing with hopelessness, physical and emotional fatigue, and sadness. Considering the ever-increasing number of people suffering from this disease, the necessity for automated depression severity assess-ment systems is profound. These systems can be used in treatment procedures, and the findings provided from learned models can help us better understand the dynamics of depression. To help in the solution to this illness, we propose a modular deep learning pipeline that uses speech transcripts as input for depression severity prediction. Through our pipeline, we investigate the role of popular deep learning archi-tectures in creating representations for depression assessment. To extend the depression assessment literature on text modality, we provide a thorough anal-ysis of sentence statistics and their effects on model training. We also present an investigation regarding the use of sentiment information for depression assess-ment. Evaluation of the proposed architectures is performed on the publicly available Extended Distress Analysis Interview Corpus dataset (E-DAIC). Through the results and discussions, we show that informative representations for depression assessment can be obtained without exploiting the temporal dynamics between sentences. Our proposed non-temporal model outperforms the state of the art by %8.8 in terms of Concordance Correlation Coefficient (CCC). In light of our findings on trained models and data statistics, we discuss how recurrent structures can have a bias toward certain sequence lengths during training and that shorter sentences can be more informative during inference. Our experimental results suggest that relying on semantic information rather than sentiment information, contrary to previous literature, may be more reliable for depression assessment. | en_US |
| dc.description.degree | M.S. | en_US |
| dc.description.statementofresponsibility | by Ergün Batuhan Kaynak | en_US |
| dc.embargo.release | 2023-03-13 | |
| dc.format.extent | xv, 66 leaves : illustrations (color), charts ; 30 cm. | en_US |
| dc.identifier.itemid | B161316 | |
| dc.identifier.uri | http://hdl.handle.net/11693/110549 | |
| dc.language.iso | English | en_US |
| dc.publisher | Bilkent University | en_US |
| dc.rights | info:eu-repo/semantics/openAccess | en_US |
| dc.subject | Depression severity assessment | en_US |
| dc.subject | Speech transcription analysis | en_US |
| dc.subject | Text analysis | en_US |
| dc.subject | Deep learning | en_US |
| dc.title | Modeling speech transcriptions for automatic assessment of depression severity | en_US |
| dc.title.alternative | Depresyon şiddeti değerlendirmesi için konuşma çevriyazılarının modellenmesi | en_US |
| dc.type | Thesis | en_US |