Namhee Kwon, Shahruk Hossain, Nate Blaylock, Henry O'Connell, Naomi Hachen, Joseph Gwin
We developed a model for detecting anxiety and depression from telephony recordings between a customer and a representative at a call center using vocal features and a deep neural network. Our binary classification model using x-vectors outperformed the use of the other acoustic features such as i-vectors and openSMILE features, as well as linguistic or text-based features. Our models were built based on self-reported scores: GAD-7 for anxiety and PHQ-8 for depression. Especially, the anxiety model’s performance is very similar to the GAD-7 score’s screening accuracy. A prior study compared self-reported GAD-7 scores to an actual mental health professional’s diagnosis of anxiety disorder and reported sensitivity and specificity of 0.74 and 0.54 respectively, and our model showed a sensitivity of 0.70 and a specificity of 0.54. This study exhibits the potential of voice analysis on topic-independent speech, particularly from 8 kHz phone conversations, to iden-tify anxiety and depression.
{"title":"Detecting Anxiety and Depression from Phone Conversations using x-vectors","authors":"Namhee Kwon, Shahruk Hossain, Nate Blaylock, Henry O'Connell, Naomi Hachen, Joseph Gwin","doi":"10.21437/smm.2022-1","DOIUrl":"https://doi.org/10.21437/smm.2022-1","url":null,"abstract":"We developed a model for detecting anxiety and depression from telephony recordings between a customer and a representative at a call center using vocal features and a deep neural network. Our binary classification model using x-vectors outperformed the use of the other acoustic features such as i-vectors and openSMILE features, as well as linguistic or text-based features. Our models were built based on self-reported scores: GAD-7 for anxiety and PHQ-8 for depression. Especially, the anxiety model’s performance is very similar to the GAD-7 score’s screening accuracy. A prior study compared self-reported GAD-7 scores to an actual mental health professional’s diagnosis of anxiety disorder and reported sensitivity and specificity of 0.74 and 0.54 respectively, and our model showed a sensitivity of 0.70 and a specificity of 0.54. This study exhibits the potential of voice analysis on topic-independent speech, particularly from 8 kHz phone conversations, to iden-tify anxiety and depression.","PeriodicalId":213932,"journal":{"name":"SMM22, Workshop on Speech, Music and Mind 2022","volume":"14 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122946042","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
I. Zubiaga, Ignacio Menchaca, Mikel de Velasco, R. Justo
Concern for mental health has increased in the last years due to its impact in people life quality and its consequential ef-fect on healthcare systems. Automatic systems that can help in the diagnosis, symptom monitoring, alarm generation etc. are an emerging technology that has provided several challenges to the scientific community. The goal of this work is to de-sign a system capable of distinguishing between healthy and depressed and/or anxious subjects, in a realistic environment, using their speech. The system is based on efficient representations of acoustic signals and text representations extracted within the self-supervised paradigm. Considering the good results achieved by using acoustic signals, another set of experiments was carried out in order to detect the specific illness. An analysis of the emotional information and its impact in the presented task is also tackled as an additional contribution.
{"title":"Mental Health Monitoring from Speech and Language","authors":"I. Zubiaga, Ignacio Menchaca, Mikel de Velasco, R. Justo","doi":"10.21437/smm.2022-3","DOIUrl":"https://doi.org/10.21437/smm.2022-3","url":null,"abstract":"Concern for mental health has increased in the last years due to its impact in people life quality and its consequential ef-fect on healthcare systems. Automatic systems that can help in the diagnosis, symptom monitoring, alarm generation etc. are an emerging technology that has provided several challenges to the scientific community. The goal of this work is to de-sign a system capable of distinguishing between healthy and depressed and/or anxious subjects, in a realistic environment, using their speech. The system is based on efficient representations of acoustic signals and text representations extracted within the self-supervised paradigm. Considering the good results achieved by using acoustic signals, another set of experiments was carried out in order to detect the specific illness. An analysis of the emotional information and its impact in the presented task is also tackled as an additional contribution.","PeriodicalId":213932,"journal":{"name":"SMM22, Workshop on Speech, Music and Mind 2022","volume":"2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125327790","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Multi-task learning from Unlabelled Data to Improve Cross Language Speech Emotion Recognition","authors":"Hafiz Shehbaz Ali, S. Latif","doi":"10.21437/smm.2022-4","DOIUrl":"https://doi.org/10.21437/smm.2022-4","url":null,"abstract":"","PeriodicalId":213932,"journal":{"name":"SMM22, Workshop on Speech, Music and Mind 2022","volume":"28 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128578291","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Edward L. Campbell, Laura Docío-Fernández, C. García-Mateo, Andre Wittenborn, J. Krajewski, N. Cummins
The Continuous Sleepiness detection task was a Sub-Challenge developed in the 2019 INTERSPEECH Computational Paralinguistics Challenge (ComParE). The associated speech corpus has been a reference in last years for the speech-based detection of sleepiness conditions. In this paper, we proposed a Sequence-to-Sequence model with global attention mechanism to accomplish this detection task. To the best of the authors’ knowledge, this is the first such an approach has been proposed for this task. Given the smaller size of this corpus, we utilise a small batch size, and augment our system with a score ensembling strategy to deliver the final decision. Despite the high complexity of our approach, it produces exceptionally competitive performances on the test-set, producing the second best performance to date. This result highlights the benefits of using deep-learning approaches, even with smaller sized speech-corpora.
{"title":"Automatic detection of short-term sleepiness state. Sequence-to-Sequence modelling with global attention mechanism.","authors":"Edward L. Campbell, Laura Docío-Fernández, C. García-Mateo, Andre Wittenborn, J. Krajewski, N. Cummins","doi":"10.21437/smm.2022-2","DOIUrl":"https://doi.org/10.21437/smm.2022-2","url":null,"abstract":"The Continuous Sleepiness detection task was a Sub-Challenge developed in the 2019 INTERSPEECH Computational Paralinguistics Challenge (ComParE). The associated speech corpus has been a reference in last years for the speech-based detection of sleepiness conditions. In this paper, we proposed a Sequence-to-Sequence model with global attention mechanism to accomplish this detection task. To the best of the authors’ knowledge, this is the first such an approach has been proposed for this task. Given the smaller size of this corpus, we utilise a small batch size, and augment our system with a score ensembling strategy to deliver the final decision. Despite the high complexity of our approach, it produces exceptionally competitive performances on the test-set, producing the second best performance to date. This result highlights the benefits of using deep-learning approaches, even with smaller sized speech-corpora.","PeriodicalId":213932,"journal":{"name":"SMM22, Workshop on Speech, Music and Mind 2022","volume":"140 1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128982152","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：